(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Proton transfer in catalysis by the carbonic anhydrases"









PROTON TRANSFER IN CATALYSIS BY THE CARBONIC ANHYDRASES 



By 
J. NICOLE EARNHARDT 



A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL 

OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT 

OF THE REQUIREMENTS FOR THE DEGREE OF 

DOCTOR OF PHILOSOPHY 

UNIVERSITY OF FLORIDA 

1998 






ACKNOWLEDGMENTS 

I would foremost like to extend my deepest appreciation to my mentor, Dr. David 
Silverman. Dr. Silverman has provided me with a productive, motivational and well-funded 
environment to conduct my research. His tutelage has given me the self-confidence to ensure 
every project I pursue is successful. I am also indebted to Dr. C. K. Tu for his countless 
hours of time and patience in teaching me instrumentation operations and his unselfish 
assistance with data analysis, experiments, troubleshooting and brainstorming. Dr. Minzhang 
Qian was invaluable to my research; she was responsible for generating most of the mutant 
forms of carbonic anhydrases which I used for my doctoral research. I want to thank Dr. 
Philip Laipis for being co-chairman of my graduate committee and for his significant 
contributions to my work and development as a scientist. I would also like to thank past and 
present technicians of the Silverman and Laipis laboratories, including Bret Schipper and Nina 
Wadhwa, as well as our collaborators from the laboratories of Richard Tashian and Ronald 
Viola. Also, I thank my friends in the Silverman lab, who have not only been helpful and 
provided sound advice but have also been a real pleasure to work with every day. I would 
like to thank Dr. Brian Cain for providing assistance during my change in laboratories in my 
third year. I would like to thank Dr. Harry Nick for taking a specific interest in my success 
as a graduate student. He has always treated me as one of his own students since I came to 
the University of Florida. I am looking forward to working with him during my postdoctoral 
studies. 



Most importantly, I would like to thank the people closest to my heart. My father, 
brother, and grandma are truly my best friends for life. I know I could never fail in anything 
I pursue with my family supporting me. To Chris: a simple thank you cannot suffice for 
someone who guided me through the roughest times of my life and graduate career. He has 
made my life everything that it should be and even more. Knowing Chris and his dear family 
has made me become a better person every day. 






• •* 

in 



TABLE OF CONTENTS 

page 

ACKNOWLEDGMENTS ii 

LIST OF TABLES vii 

LIST OF FIGURES viii 

ABSTRACT x 

CHAPTERS 

1. INTRODUCTION 1 

The Mammalian Carbonic Anhydrases 1 

Physiological Function 1 

Structure 3 

Inhibition 7 

Overview of the Catalysis 8 

Interconversion of C0 2 and HC0 3 " 9 

Proton Transfer 11 

Bronsted Analysis 14 

2. THE CATALYTIC PROPERTIES OF MURINE CARBONIC 

ANHYDRASE VII 18 

Introduction 18 

Materials and Methods 20 

Expression and Purification of a Recombinant Murine CA VII cDNA . . 20 

Subcloning and Site-Directed Mutagenesis of CA VII 23 

Enzyme Purity 26 

Stopped-Flow Spectrophotometry 26 

18 Exchange 27 

Inhibition 29 

Hydrolysis of 4-Nitrophenyl Acetate 31 

Results 31 

Recombinant Murine CA VII 31 



IV 



Catalytic Activity 32 

Discussion 40 

Comparison of Isozymes of Carbonic Anhydrase 40 

Proton Transfer in CA VII 42 

Assignment of pK a Values 46 

Inhibition 49 

Conclusions 49 

3. INTRAMOLECULAR PROTON TRANSFER FROM MULTIPLE SITES IN 

CATALYSIS BY MURINE CARBONIC ANHYDRASE V 51 

Introduction 51 

Materials and Methods 53 

Site-Specific Mutagenesis 53 

Expression and Purification 55 

Stopped-flow Spectrophotometry 56 

18 Exchange 56 

Results 56 

Discussion 66 

4 CATALYSIS BY MURINE CARBONIC ANHYDRASE V IS ENHANCED 

BY EXTERNAL PROTON DONORS 75 

Introduction 75 

Materials and Methods 77 

Site-Specific Mutagenesis, Protein Expression and Purification 77 

18 Exchange 78 

Results 78 

Method 1 : Saturation Effect of Buffers on R H 2c/[E] 78 

Method 2: pH Dependence of R H2C /[E] 81 

pK a of the Donors and Acceptors Listed in Table 4-1 86 

Determination of k B Values of Table 4-1 by Method 1 88 

Solvent Hydrogen Isotope Effects 89 

Discussion 89 

Choice of Mutant and Buffers 89 

Enhancement of Catalysis 90 

Brensted Analysis 91 

Marcus Rate Theory 94 

Conclusions 97 

5. DISCUSSION, CONCLUSIONS AND FUTURE DIRECTIONS 99 

Isozyme VII 99 

Isozyme V 102 

Buffers in Catalysis 103 

Conclusions 104 



Future Work 105 

REFERENCES 108 

BIOGRAPHICAL SKETCH 114 



VI 






LIST OF TABLES 



Table page 

1-1 Cellular Location and Predominant Tissue Distribution of the Mammalian 

Carbonic Anhydrases 2 

1-2 Maximal Values of the Steady- State Constants with Values of Apparent 

pK a for C0 2 Hydration Catalyzed by the Mammalian Carbonic Anhydrases. ... 10 

2-1 Inhibition Constants Kj (Nanomolar) for Isozymes of Carbonic Anhydrase 

Determined by 18 Exchange 30 

2-2 Catalysis by Carbonic Anhydrase VII: Maximal Values of Steady-State 

Constants with Values of Apparent pK a 34 

3-1 Maximal Values of k^/Kn, and k cal for C0 2 Hydration and k B with pK a Values 
Obtained from their pH Profiles for Wild-type and Mutants of Murine 
Carbonic Anhydrase V 57 

3-2 Catalysis Enhanced by Proton Donors from Solution in Mutants of MCA V: 
Values of R H2C /[E], the Rate of Release of 18 0-labeled Water from the 
Active Site 65 

4-1 Constants of Equations 4-1 and 4-2 for the Enhancement by Buffers of 

Catalysis by Y64A/F65A MCA V 80 

4-2 Maximal Values of k B with pK a Obtained from their pH Profiles for the 

Y64A/F65A Mutant of Murine Carbonic Anhydrase V, in the Absence and 
Presence of Buffer: Method 2 85 

4-3 Marcus Theory Parameters for Proton Transfer in Isozymes of Carbonic 

Anhydrase 96 



vu 



LIST OF FIGURES 

Figure page 

1-1 A ribbon model of human carbonic anhydrase II determined from the 

crystal structure 4 

1-2 Residues near the active site of human carbonic anhydrase II 6 

2-1 The sequence of MCA VII cDNA and derived amino acids 21 

2-2 The pH dependence of k,.,/!^ for hydration of C0 2 and dehydration of HC0 3 " 

catalyzed by rMCA VII at 25 °C 33 

2-3 The turnover number k^, for hydration of C0 2 and dehydration of HC0 3 " 
catalyzed by rMCA VII and hydration of C0 2 catalyzed by rMCA VII H64A 
obtained by stopped-flow spectrophotometry at 25 °C 36 

2-4 Variation with pH of R H2C /[E], the proton-transfer dependent rate constant 
for the release from the enzyme of 18 0-labeled water, catalyzed by rMCA VII 
and the H64A mutant of rMCA VII 38 

2-5 The variation with pH of R H2C /[E] catalyzed by full-length rMCA VII, the 
H64A mutant of full-length rMCA VII, the truncated MCA Vllb, and a 
H64A mutant of the truncated MCA Vllb 39 

2-6 The pH dependence of k^/K^ for the hydrolysis of 4-nitrophenyl acetate 

catalyzed by rMCA VII 41 

3-1 The location of ionizable residues near the active site cavity of murine carbonic 

anhydrase V from the crystal structure 54 

3-2 The pH dependence of k^/K^, for hydration of C0 2 determined by 18 
exchange catalyzed by wild-type MCA V; K91A/Y131A MCA V; and 
Y64A/K91A/Y131AMCAVat25 °C 58 



Vlll 



3-3 The pH dependence of k cal for hydration of C0 2 determined by stopped-flow 
spectrophotometry catalyzed by wild-type; K91A/Y131A MCA V; and 
Y64A/K91A/Y131AMCAVat25 °C 60 

3-4 The pH dependence of R H2C /[E], the rate constant for release of 18 0-labeled 
water from the enzyme, catalyzed by wild-type MCA V and the mutant 
K91A/Y131AMCAVat25 °C 62 

3-5 The dependence of R H2C /[E] and R,/[E] as a function of the concentration of 

imidazole at pH 6.3 and 25 °C 64 

4-1 The dependence of R H2C /[E] and R/fE] as a function of the concentration of 

protonated 3,5-dimethyl pyridine at pH 6.3 and 25 °C 79 

4-2 Variation with pH of R H20 /[E], the proton-transfer dependent rate constant 
for the release from the enzyme of 18 0-labeled water, catalyzed by MCA V 
Y64A/F65A in the absence of buffer and in the presence of 100 mM 
3,5-dimethyl pyridine and 100 mM imidazole at 25 °C 82 

4-3 The difference in R H 2c/[E], tne proton-transfer dependent rate constant for the 
release from the enzyme of 18 0-labeled water, between MCA V Y64A/F65A 
in the absence and presence of 100 mM 3,5-dimethyl pyridine and 100 mM 
imidazole at 25 °C 83 

4-4 Dependence of the logarithm of k B (s"') on ApK a (the pK a of the zinc-bound 

water subtracted from the pK a of the donor group) 93 



IX 



Abstract of Thesis Presented to the Graduate School 
of the University of Florida in Partial Fulfillment of the 
Requirements for the Degree of Doctor of Philosophy 



PROTON TRANSFER IN CATALYSIS BY THE CARBONIC ANHYDRASES 

By 

J. Nicole Earnhardt 
December 1998 

Chairperson: Dr. David N. Silverman 

Major Department: Biochemistry and Molecular Biology 

Carbonic anhydrase (CA) catalyzes the reversible hydration of carbon dioxide to 

bicarbonate and a proton. This reaction requires proton transfer between the zinc-bound 

water at the active site and aqueous solution for each cycle of catalysis, and the most efficient 

isozymes facilitate this transfer by the proton shuttle residue, histidine 64. In this work, I 

have characterized the catalytic properties of a recombinant murine carbonic anhydrase VII 

(CA VII), using stopped-flow spectrophotometry and 18 exchange measured by mass 

spectrometry. CA VII has steady-state constants similar to the most efficient isozymes of 

carbonic anhydrase, CA II and IV, and is strongly inhibited by the sulfonamides 









ethoxzolamide and acetazolamide. The magnitude of k caI , near 10 6 s" 1 , its pH profile, 18 0- 
exchange data for both wild-type and a histidine 64 to alanine mutant, and inhibition by 
CuS0 4 all suggest that His 64 is a proton transfer group in CA VII. A truncation mutant of 
CA VTI, in which 23 residues from the amino-terminal end were deleted, has its rate constant 
for intramolecular proton transfer decreased by an order of magnitude compared with the full 
length counterpart. There is no change in k^/K^ which measures the interconversion of C0 2 
and HC0 3 ~ in a stage of catalysis that is separate and distinct from the proton transfers. This 
is the first evidence supporting a role of the amino-terminal end in enhancing proton transfer 
by any CA. 

These studies of CA VII showed proton transfer capability even in the absence of His 
64. The best example of proton transfer by a CA not having a histidine at position 64 is the 
mitochondrial isozyme, murine CA V (MCA V). MCA V has a sterically constrained tyrosine 
at position 64; it is not an effective proton shuttle, yet catalysis still achieves a maximal 
turnover in C0 2 hydration of 3 x 10 5 s" 1 at pH > 9. This study has identified several basic 
residues, including Lys 91 and Tyr 131, located near the mouth of the active-site cavity that 
contributes to proton transfer. Comparison of k cal in catalysis upon replacement of Lys 91 
and Tyr 1 3 1 with alanine yielded a reduction in catalytic activity of 50% from wild-type. The 
corresponding double mutant showed a strong antagonistic interaction between these sites 
suggesting a cooperative behavior in facilitating the proton transfer step of catalysis. These 
replacements caused relatively small changes in k^/K,,, indicating that the replacements of 
proton shuttles have not caused structural changes that affect reactivity at the zinc. 






XI 






A wider range of pK a values for the proton donors was achieved by measuring 
enhancement of catalysis by imidazole, pyridine, and morpholine buffers in solution. These 
buffers enhance proton transfer steps in a mutant of MCA V, Y64A/F65A, to values similar 
to that of the intramolecular counterpart, histidine, in the mutant Y64H/F65A. The rate 
constants for proton transfer from these buffers to the zinc-bound hydroxide in catalysis of 
MCA V show a direct correlation to the difference in acid and base strength of the catalysts. 
Application of Marcus rate theory shows that this proton transfer has a small intrinsic energy 
barrier (near 0.8 kcal/mol), which is also characteristic of nonenzymic rapid proton transfer 
between nitrogen and oxygen acids and bases in solution. The Marcus parameters yield a 
large thermodynamic component (near 10 kcal/mol). This work is a counterpart to studies 
of proton transfer involving histidine 64 and identifies solvent and active-site reorganization 
as a dominant feature in proton transfer in catalysis by carbonic anhydrase. 



xn 












CHAPTER 1 
INTRODUCTION 



The Mammalian Carbonic Anhydrases 

Carbonic anhydrases are zinc metalloenzymes that catalyze the reversible hydration 
of carbon dioxide to form bicarbonate and a proton: 

C0 2 + H 2 ** HCO/ + H + 
Carbonic anhydrase (CA) from animals, plants, archaebacteria and eubacteria have been 
classified into three gene families based on unrelated evolutionary histories. These are 
designated a, P, and y (Hewett-Emmett and Tashian, 1996). The mammalian carbonic 
anhydrases are all included in the a-class and will be discussed in detail in the following pages. 
Physiological Function 

There are at least seven functional mammalian isozymes of CA, referred to as CA I 
through CA VII. The different isozymes are found in various locations in the cell. Isozymes 
I, II, HI, and VII are all found in the cytosol. CA IV is a glycosylphosphatidylinositol (GPI)- 
anchored membrane protein, CA V is found in the mitochondria and CA VI is a secretory 
protein (Table 1-1, reviewed in Dodgson, 1991 and Sly and Hu, 1995). 

The isozymes of CA are distributed throughout the different body tissues and serve 
various physiological functions, all of which are associated with the reversible hydration of 
C0 2 (reviewed in Dodgson, 1991 and Sly and Hu, 1995). For example, isozyme II, 

1 



Table 1-1: Cellular Location and Predominant Tissue Distribution of the Mammalian Carbonic 
Anhydrases. 



Isozyme 


Cellular Location 


Main Tissue Distribution 


CAI 


Cytosol 


Red cells 


CAII 


Cytosol 


Ubiquitous, red cells and secretory tissues 


CAIII 


Cytosol 


Skeletal muscle and adipose tissue 


CAIV 


Membrane-bound 


Ubiquitous, lung, and kidney 


CAV 


Mitochondrial 


Liver and kidney 


CAVI 


Secretory 


Salivary glands 


CAVII 


Cytosol 


Salivary gland", brain b , lung c 


Note: This information is reviewed in 


Dodgson, 1991 and Sly and Hu, 1995. 



a Montgomery et al., 1991. 
b Lakkis et al., 1997. 
c Lingetal., 1994. 



having the most diverse tissue distribution of all the carbonic anhydrases, provides such 
physiological functions as H + production for renal acidification of urine and gastric acid 
secretion, and HC0 3 " for the production of pancreatic juice, saliva, ocular and cerebrospinal 
fluid. The involvement of isozyme II in the formation of aqueous humor has made this CA 
a target enzyme for the treatment of glaucoma (Maren, 1997). Sulfonamide inhibitors of CA 
are used to reduce the production of ocular fluid and thereby decrease intraocular pressure 
that is a condition of this disease. CA II is also involved in C0 2 transport and/or exchange 
in the kidney, red cells and lung and CA IV provides this same function in lung, brain, 
skeletal and heart muscles. There is facilitated diffusion of C0 2 in skeletal muscle by isozyme 
HI, HC0 3 " reabsorption in the kidney by isozyme IV, and pH regulation by isozyme VI in the 
saliva. Isozyme V has a metabolic function of providing HC0 3 ' to carbamoyl phosphate 
synthetase I and pyruvate carboxylase for ureagenesis and gluconeogenesis, respectively. CA 
V has also been postulated to have a role in lipogenesis (Dodgson, 1991). 
Structure 

Of the seven isozymes, crystal structures of human CA I (Kannan et al., 1975), II 
(Eriksson et al., 1988a; Hakansson et al., 1992), bovine III (Erickson and Liljas, 1993), 
human and murine IV (Stams et al., 1996; 1998) and murine V (Boriack-Sjodin et al., 1995) 
have been determined. All these isozymes have common features including nearly identical 
structures around the active site by comparison of their backbone atoms. A representative 
structure of human CA II is shown in Figure 1-1 . Each isozyme is a spherical molecule with 
a 24 residue amino-terminal tail that is loosely fit in relation to the rest of the molecule as 
determined in crystal structures of isozymes I, II, III and IV (the CA V crystal is of an amino- 
















Figure 1-1. A ribbon model of human carbonic anhydrase II determined from the crystal 
structure (Hakansson et al., 1992). The three histidine ligands of the zinc are 94, 96, and 
119. The intramolecular proton shuttle residue, His 64, is indicated. 



terminal truncated protein). These isozymes are structurally dominated by a central 10- 
stranded twisting P-sheet with short helices located on the surface of the molecule. 

The active site is located at the bottom of a central cavity approximately 15 A wide 
at the surface and 15 A deep. In the active site, a divalent zinc ion is coordinated to the 
nitrogen atoms of three histidine residues and a H 2 molecule (Figure 1-1). The histidine 
ligands are held in a strict tetrahedral geometry through hydrogen bonds to a series of 
second-shell residues referred to as indirect ligands (Christianson and Fierke, 1996). The 
histidine and indirect ligands are conserved throughout the functional mammalian carbonic 
anhydrases with the exception of the indirect ligand, 244, which is hydrogen bonded to a 
histidine ligand by its main chain carbonyl group. 

Another feature of the mammalian carbonic anhydrases that is strictly conserved and 
observed in each crystal structure is the hydrogen bonding system of two amino acid residues, 
Thr 199 and Glu 106, with the zinc-bound hydroxide (Figure 1-2). In this system the zinc- 
bound hydroxide is a hydrogen-bond donor to the side-chain hydroxyl of Thr 199, which in 
turn, is a hydrogen-bond donor to the carboxylate of Glu 106. The zinc-bond hydroxide also 
forms a hydrogen bond with another water molecule (referred to as the "deep water") located 
in a hydrophobic pocket in the active site and is hydrogen bonded to the amido group of Thr 
199. This system apparently restricts the orientation of the zinc-bound hydroxide for efficient 
reaction with substrate C0 2 (Merz, 1990; Krebs et al., 1993a; Liang et al., 1993a) and is 
important for binding of bicarbonate (Xue et al., 1993), sulfonamides and many anionic 
inhibitors (Erickson et al., 1988b; reviewed in Liljas et al., 1994). 


















Glul06-< 



His 96 




Hydrophobic 
pocket 



His 94 



Figure 1-2. Residues near the active site of human carbonic anhydrase II. The three histidine 
ligands of the zinc are 94, 96, and 119. The two amino acids comprising the hydrogen- 
bonded system are, Glu 106 and Thr 199. The line at the right indicates the area of the 
hydrophobic pocket. 



7 
There is a hydrophobic and a hydrophilic area located towards the surface of the 
active site cavity and the difference in amino acids found in these regions result in the 
variations in the properties of the seven CA isozymes. Several studies suggest that C0 2 is 
weakly bound in the active site and interacts with the hydrophobic area; yet, there is no crystal 
structure with detectable bound C0 2 for verification nor has any one experiment specifically 
located the C0 2 binding site (Figure 1-2; Krebs et al., 1993b). The residues that comprise this 
hydrophobic site are Val 121, Val 143, Leu 198 and Trp 209 in isozyme II (Alexander et al., 
1991; Fierke et al., 1991). This hydrophobic cavity is also important for inhibitor binding as 
discussed below. 

Last, in the active site of isozyme II there is an ordered array of hydrogen-bonded 
water molecules detectable through X-ray crystallography (Erickson et al., 1988a, Hakansson 
et al., 1992). This hydrogen-bonded water chain is necessary for proton transfer from the 
zinc-bound water to buffer in solution which can occur through a proton shuttle residue such 
as histidine 64 in isozyme II (Figure 1-1; Venkatasubban and Silverman, 1980; Tu et al., 
1989a). This will be discussed in future sections. 
Inhibition 

Inhibitors of CA include a variety of anions, neutral organic molecules, sulfonamides, 
and metal ions such as Cu(II) and Hg(II) (reviewed in Liljas et al., 1994). Competitive 
inhibitors with respect to C0 2 in steady-state experiments include phenol in isozyme II and 
imidazole in isozyme I. Small anion inhibitors include, among many others, azide, 
thiocyanate, nitrate and formate. All inhibitors, with the exception of the dipositive metal ions 
mentioned above, are found to bind at or near the zinc ion within the hydrogen-bonded 



8 
system of the zinc-bound hydroxide, Glu 106 and Thr 199 (Figure 1-2). Almost all inhibitors 
appear to displace the deep water, most coordinate with the metal ion, and many contribute 
a hydrogen bond to the hydroxyl of Thr 199. 

All aromatic and certain heterocyclic sulfonamides inhibit catalysis by binding to the 
zinc ion as anions with the nitrogen atom of the sulfonamide group, R-S0 2 -NH" (reviewed 
in Liljas et al., 1994). More specifically, the NH" group replaces the zinc-bound water 
molecule and hydrogen bonds to the hydroxyl group of Thr 199 (Figure 1-2; Erickson et al., 
1988b). One of the sulfonamide group oxygen atoms forms a hydrogen bond with the peptide 
NH of Thr 199 while displacing the deep water molecule and the second sulfonamide oxygen 
has no contact and is pointing away from the zinc. The sulfonamide is positioned in the 
hydrophobic pocket with van der Waals contacts to the hydrophobic residues and these 
interactions all depend on the substitutions in the aromatic ring of the sulfonamide. These 
hydrophobic residues lead to variations in the inhibitory properties of the various isozymes 
with respect to the sulfonamides. 

Metal ions inhibit CA, especially the isozymes with histidine functioning as a proton 
shuttle. Hg 2 * and Cu 2+ bind to the proton shuttle residue, His 64, in CA II and prevent proton 
transfer from this site (Tu et al., 1981). Both nitrogens of the histidine are found to bind 
mercury at half occupancy in the crystal structure of isozyme II (Erickson et al., 1988b). 

Overview of the Catalysis 

The enzymatic mechanism has been well studied in most of the mammalian carbonic 
anhydrases. The catalysis of the reversible hydration of carbon dioxide to form bicarbonate 



9 
and a proton occurs in two separate and distinct stages and is therefore referred to as a ping 
pong mechanism. The first stage of catalysis comprises the hydration of C0 2 involving the 
direct nucleophilic attack of zinc-bound hydroxide on substrate CO, to yield bicarbonate, the 
departure of bicarbonate leaves a water bound to the zinc (equation 1-1). The second stage 
requires a proton to be transferred from the zinc-bound water to buffer in solution (designated 
as B in equation 1-2) to regenerate the zinc-bound hydroxide (H*E indicates a protonated 
shuttle residue; Christianson and Fierke, 1996; Lindskog, 1997). 

+ H 2 
C0 2 + EZnOH" ** EZnHCCV ** HCCV + EZnH 2 ( 1 - 1 ) 

EZnH 2 + B ** H + EZnOH" + B * EZnOH" + BH + (1-2) 

Interconversion of CQ 2 and HCQ 3 " 

The steady state ratio k^/K,, contains rate constants for the steps from the binding 
of CO, up to and including the first irreversible step in catalysis, which is the departure of 
HC0 3 ' (Silverman and Lindskog, 1988). The carbonic anhydrases rapidly catalyze the 
conversion of C0 2 to HC0 3 ' as represented in the magnitude of the steady state rate constant 
k^/Kn, (Table 1-2). For isozyme II, the maximal value of k cat /K„, approaches that for the 
diffusion-controlled limit for the encounter rate of enzyme and substrate which is estimated 
at 10 9 to 10 10 MV , suggesting that catalysis occurs as fast as C0 2 can diffuse into the active 
site (Khalifah, 1973; Lindskog and Coleman, 1973). 

For most of the mammalian carbonic anhydrases, the pH profile of k^/K^ is described 
by a single ionization with a pK a near 7 with maximum activity at high pH. CA III is an 






10 



Table 1-2: Maximal Values of the Steady-State Constants with Values of Apparent pK a for 
C0 2 Hydration Catalyzed by the Mammalian Carbonic Anhydrases. 

Isozyme k^ (x lO^MV) pK,^^ k ca , (x IP" 5 s" 1 ) pK^,, 

Human CAI a 5 7.0 b 2 

Human CAII a 15 7.1 14 7.1 

Human CAIII C 0.03 <6.0 0.1 >8.5 

Murine CAIV d 3.2 6.6 11 6.3 and 9.1 






Murine CAV e 3 7.4 3 9.2 

RatCAVI f 1.6 -- 0.7 

Murine CA VII 8 7.6 6.2 and 7.5 9.4 6.2 and 8.2 

Note : Data obtained at 25 °C. 

a Khalifah, 1971. pK^^ is not available because k,.,, is reported to increase with pH without 
reaching a plateau up to pH 8.7. 

b Behravan et al., 1991. pK a(kcal/Km) was determined from the pH dependence of the second- 
order rate constant for the catalyzed hydrolysis of 4-nitrophenyl acetate. 
c Jewell et al., 1991. 
d Hurtetal., 1997 
'Hecketal., 1994. 

f Feldstein and Silverman, 1984. Data obtained at pH 7.5, and probably does not represent 
maximal values. Therefore, values of pK a are not yet determined. 
8 Data obtained from this work (Chapter 2). 



11 

exception in that k ca /K m is described by a single ionization with a pK, near 5 (Table 1-2). This 
pK a found for k^/K^ is associated with the ionization state of the zinc-bound water. 
Proton Transfer 

The proton transfer step in the hydration direction of catalysis proceeds from the zinc- 
bound water to buffer in solution (equation 1-2). Both intramolecular proton transfer to a 
shuttle residue, such as His 64 in CA II, and intermolecular proton transfer to small buffers 
that fit in the active site occur through a hydrogen-bonded water structure that surrounds the 
metal in the active site cavity. A network of hydrogen-bonded waters is observed in the 
crystal structure of CA II, and at least two intervening water molecules are found between 
the imidazole ring of the proton shuttle, His 64, and the zinc-bound water (Erickson et al., 
1988a). Consistent kinetic evidence for proton transfer along this network of hydrogen- 
bonded waters has been obtained by studying the solvent hydrogen isotope effects on k cat for 
the hydration of C0 2 in C A II. Here, two or more protons are determined to be in motion 
during intramolecular proton transfer (Venkatasubban and Silverman, 1980). Using site- 
directed mutagenesis in isozyme II to place bulky residues in the active site at position 65 
reduces proton transfer in catalysis and provides evidence for the disruption of this water 
structure (Jackman et al., 1996; Scolnick and Christianson, 1996). Proton transfer between 
water molecules is described by the "Grotthus chain mechanism" (Agmon, 1995). This 
mechanism refers to sequential proton transfer steps between water molecules which does not 
require the same proton to be transferred along the water chain. In addition, for the carbonic 
anhydrases, the first proton transfer from the zinc-bound water to the next adjacent water 






12 
molecule is postulated to form a hydronium-like ion and determines the rate of CA catalysis 
(Liang and Lipscomb, 1988). 

Intramolecular proton transfer. The steady state constant k cal contains the rate 
constants from the enzyme-substrate complex to the end of catalysis; in CA this includes the 
proton transfer step to regenerate the zinc bound water (Silverman and Lindskog, 1988). 
Among the carbonic anhydrases, isozyme II has the greatest turnover number of 10 6 s' 1 (Table 
1-2; Khalifah, 1971). pH profiles of k cat for this isozyme follow a titration curve with a pK a 
of 7, therefore, the catalytic rate, k ca „ is dependent upon the ionization of a residue, or 
residues, with pK a 's near 7. The high rate of catalysis of isozyme II and the ionization 
observed in pH profiles of k^ is attributed to a histidine at position 64 that has been identified 
as the catalytic residue for intramolecular proton transfer (Tu et al., 1989a). Studies using 
isotope effects, pH dependencies, and chemical rescue have shown that these intramolecular 
proton transfer steps are rate-determining for maximal velocity (Silverman and Lindskog, 
1988; Tu et al., 1989a). Therefore, the high rate of catalysis of the hydration of C0 2 is 
determined almost entirely by the intramolecular proton transfer between His 64 and the zinc- 
bound water (Lindskog, 1984; Rowlett, 1984). 

The role of position 64 in catalysis by carbonic anhydrase II has been well studied. 
The efficiency of His 64 as an acid-base catalyst in isozyme II is attributed to its pK a value of 
7 which is similar to that of the zinc-bound water, and to optimal spatial location and 
environment in the active site. The location of the imidazole side chain of His 64 is 7 A from 
the zinc and it extends into the active-site cavity with no apparent interactions with other 
residues (Erickson et al., 1988a). 



13 
Among the other seven isozymes, CA IV and CA VII both have a histidine at position 
64 and appropriately yield rate constants for proton transfer approaching that of the high 
efficiency isozyme II (Table 1-2; Hurt et al., 1997, Earnhardt et al., 1998a). However, it must 
be noted that His 64 is not the sole proton shuttle residue in these two isozymes as indicated 
by two ionizations in the pH profiles of k^ (Table 1-2; discussed further in Chapter 2). In CA 
I a histidine at position 64 is present, however, it is not functioning as a proton shuttle 
(Behravan et al., 1991). Isozyme III has the slowest turnover number of 10 4 s' 1 for the CA 
isozymes. The corresponding residue at position 64 in CA III, lysine, lacks significant 
intramolecular proton transfer capability and the proton shuttles residue(s) representing the 
ionization in k^, remain unknown (Table 1-2; Jewell et al., 1991). However, in mutants of 
isozyme III where a histidine has been inserted at position 64 and 67, the maximal k cat values 
are restored to values closer to isozyme II (Jewel et al., 1991, Ren et al., 1995). Similar 
observations have been found for isozyme V with a tyrosine at position 64, and as a result 
turnover numbers do not approach those of isozyme II (discussed in Chapter 3, Heck et al., 
1994; Earnhardt et al., 1998b). However, upon site-directed mutagenesis, isozyme II like 
properties are found in a Y64H/F65A mutant of isozyme V (Heck et al., 1996). 

Intermolecular proton transfer . As described above, for the hydration of C0 2 
catalyzed by C A, buffer in solution is the final acceptor of protons that are transferred from 
the zinc-bound water. A buffer-dependent step in catalysis at low buffer concentrations is 
observed when the catalyzed initial velocity of C0 2 hydration is determined at buffer 
concentration less than 10 mM for isozyme II (Silverman and Tu, 1975; Jonsson et al., 1976). 
By contrast, intramolecular proton transfer is found to be rate limiting at high buffer 
concentrations. 



14 
Buffer-mediated enhancements are observed in k^, for hydration of CO, by CA II and 
are associated with the intermolecular proton transfer from the active site to solution. For 
the proton independent steps of equation 1-1, no enhancements are observed in k^/K^, upon 
addition of buffer. Under steady-state conditions in isozyme II, the rate constants for proton 
transfer from the enzyme to buffer in the catalyzed hydration of C0 2 depend on the difference 
in pK a between the enzyme as proton donor and the buffer as acceptor, consistent with 
bimolecular proton transfer between nitrogen and oxygen acids and bases in solution (Rowlett 
and Silverman, 1982). In this work, the absence of a trend in the structure of buffers that 
transfer protons yields supporting evidence for proton acceptance from a shuttle residue on 
the enzyme, which is known to be His 64 in CA II, instead of proton transfer directly with the 
zinc-bound water. 

Studies of buffer enhancement under chemical equilibrium conditions in isozymes II 
and III have also demonstrated that small buffers that fit in the active site can provide an 
intermolecular proton shuttle group from the zinc-bound water to solution by directly 
accepting protons through intervening water bridges (Silverman and Tu, 1975; Tu et al., 
1990). Buffers that lead to enhancement include imidazole, pyridine, and morpholine buffers 
and their derivatives and the observed enhancements in catalysis depend on the concentration 
of buffer and the buffer's chemical properties. 
Bronsted Analysis 

As discussed in the previous section, Rowlett and Silverman (1982) have correlated 
rate constants for intermolecular proton transfer to the difference in acid and base strength 
of the catalysts under steady state conditions, which is referred to as a Bronsted analysis. 



15 
Bronsted plots for intramolecular proton transfer in isozyme III have also been constructed. 
Using chemical equilibrium methods, the rate constants for proton transfer were determined 
from a series of mutants with a histidine or glutamates and aspartates as proton shuttles 
placed at position 64 (Silverman et al., 1993; Tu et al., 1998) and position 67 (Ren et al., 
1995). Variations in the pK a of the acceptor group under these conditions, the zinc-bound 
water, were obtained by mutagenesis of an active site residue, Phe 198 to either Leu or Asp 
(LoGrasso et al., 1991, 1993). An increase upon introduction of these residues in pK a of the 
zinc-bound water is possibly due to a change in the interaction of the hydroxyl side-chain of 
Thr 199 with the zinc-bound water that is transmitted through these mutants (Chen et al., 
1993). 

The resulting Brensted curves could be fit by the Marcus rate theory. This allows the 
energy required for proton transfer from either position 64 or 67 in these experiments to be 
determined in terms of the intrinsic kinetic barrier for proton transfer, which is found to be 
consistently low (1.3 to 2.2 kcal/mol; Silverman et al., 1993, Ren et al., 1995; Tu et al., 
1998). Solvent hydrogen isotope effects studied with small buffers in human CA II under 
steady-state conditions also exhibit a low intrinsic kinetic barrier to proton transfer (~1 
kcal/mol, Taoka et al., 1994). This low intrinsic kinetic barrier defines the energy required 
for proton transfer between nitrogen and oxygen acids and bases in CA and is similar in 
magnitude to the intrinsic kinetic barrier for bimolecular proton transfer between these two 
groups in solution (2 kcal/mol; Silverman et al., 1993). However, the energy required to 
orient the protein and/or active site water for this efficient proton transfer, expressed as a 
work function in Marcus theory, is large, 10 kcal/mol, and accounts for the slow rate of CA 



16 
catalysis (10 6 s" 1 ) when compared to that of proton transfer in excited states (10 12 s' 1 ) 
(Silverman et al., 1993; Ren et al., 1995). Proton transfer from the zinc-bound water to the 
proton shuttle group occurs through an active site water structure. Therefore, water is 
essential for proton transfer and may be involved in the efficiency of the carbonic anhydrases 
due to a required reorganization of the water lattice before proton transfer can occur. This 
type of analysis will be discussed in detail in work on isozyme V in Chapter 4. 

In the coming chapters, a detailed description will be provided of the catalytic 
properties of a new isozyme of carbonic anhydrase, isozyme VII. This work will establish the 
intramolecular proton shuttle residue as histidine 64 in this isozyme and provide novel insight 
into the influence of the amino-terminus on proton transfer. Also, in future chapters the 
intramolecular proton shuttle residues will be identified that contribute to catalysis by the 
mitochondrial CA, isozyme V. These proton shuttles are located at more distant sites from 
the zinc than position 64. Last, chemical rescue experiments will be described that involve 
buffers in solution as intermolecular proton shuttles. Small imidazole, morpholine and 
pyridine type buffers enhance catalysis in isozyme V, an isozyme that lacks a single 
predominant proton shuttle such as histidine 64. Overall, this work will investigate 
intramolecular proton transfer from near and distant proton shuttles in isozymes V and VII 
and intermolecular proton transfer from small buffers that lacks any distance requirements in 
isozyme V. This discussion of carbonic anhydrase describes proton transfer in a very well- 
defined and accessible system. It is hoped that the results obtained in this study will be 
applicable to describe proton transfers in much more complex systems that almost surely 
involve proton transfer through intervening water molecules. These complex proton 



17 
translocation systems that are under intense current scrutiny include rhodopsin of visual 
pigment, cytochrome oxidase in cellular metabolism, and the photosynthetic reaction center 
of plants. 












CHAPTER 2 
THE CATALYTIC PROPERTIES OF MURINE CARBONIC ANHYDRASE VII 



Introduction 

Carbonic anhydrase VII has recently been discovered by gene isolation from a human 
genomic library using a mouse CA II cDNA clone as a probe (Montgomery et al., 1991). The 
gene structure of CA VII is found to be very similar to the other functional isozymes, and CA 
VII is postulated to be a cytosolic enzyme (Montgomery et al., 1991). Phylogenetic analysis 
based on the amino acid sequences of the carbonic anhydrase isozymes closely relates CA VII 
to the mitochondrial CA V (Hewett-Emmett and Tashian, 1996). Both CA V and CA VII 
genes map to human chromosome 16 (Montgomery et al., 1991; Nagao et al., 1993). CA VII 
mRNA has been detected in baboon salivary gland (Montgomery et al., 1991) and rat lung 
(Ling et al., 1994). A recent detailed in situ hybridization study of CA VII mRNA expression 
in adult mouse brain revealed a wide, nonspecific distribution in different regions of the 
cerebrum and cerebellum (Lakkis et al., 1997). CA VII has also been reported from a cDNA 
library prepared from multiple sclerosis lesions found in a human patient (GenBank Ace. No. 
N78377). 

A nearly complete mouse CA VII cDNA obtained by RT-PCR using RNA isolated 
from adult mouse (C57/BL6) brain showed a protein sequence identity of about 95% with the 
human sequence; the nucleotide sequences are about 91% identical (Ling et al., 1995). It has 

18 



19 
been suggested that this close conservation in two mammalian forms of CA VII is indicative 
of the functional importance of this isozyme (Lakkis et al., 1996). This evolutionary 
conservation, together with the fact that CA VII is seemingly expressed in a wide variety of 
tissues, albeit at low levels, suggests that it may have an important general function in most 
cells. For this reason, a study of its kinetic properties is well worth investigating. 

Initial kinetic characterization of the expressed recombinant murine CA VII 
demonstrated an isozyme with rather low C0 2 hydration activity between pH 6.5 and 8.2 
when compared to bovine CA II (Lakkis et al., 1996). This chapter contains a more complete 
kinetic characterization of CA VII determined by stopped-flow spectrophotometry, 18 
exchange using mass spectrometry at chemical equilibrium, hydrolysis of 4-nitrophenyl 
acetate, and inhibition by two sulfonamides and CuS0 4 . The results of these studies show a 
high activity enzyme with the maximal values of k^/^ and k cal for C0 2 hydration 
approaching that of CA II, placing it in the subset of rapidly acting carbonic anhydrases that 
includes isozymes II and IV. The role of His 64 as a prominent proton shuttle in CA VII as 
in isozymes II and IV was verified upon analysis of the kinetic properties of a His 64 to Ala 
mutant of CA VII. Similar to murine CA IV (Hurt et al., 1997), evidence indicates that CA 
VII shows multiple intramolecular proton transfers involving the zinc-bound water and at 
least two residues that act as proton shuttles, one of which is His 64. A truncation mutant 
of CA VII lacking 23 residues at the amino-terminal end showed intramolecular proton 
transfer decreased by an order of magnitude while k^/K,,, was unchanged, suggesting a role 
for the amino-terminal end in proton transfer to the active site. And finally, CA VII was found 
to have the strongest inhibition by the sulfonamides acetazolamide and ethoxzolamide for any 
mammalian carbonic anhydrase. 



20 
Materials and Methods 

Expression and Purification of a Recombinant Murine CA VII cDNA 

I received from the laboratory of Dr. Richard E. Tashian (University of Michigan) an 
almost entire recombinant murine CA VII cDNA (rMCA VII 1 ). In Dr. Tashian's lab, Dr. 
Maha M. Lakkis obtained this cDNA by RT-PCR using RNA isolated from adult mouse 
(C57/BL6) brain (Lakkis et al., 1996). This PCR fragment was amplified using a human CA 
VII 5' primer and a mouse CA VII 3' primer (see Figure 2-1). It was then cloned into the 
glutathione S-transferase expression vector, pGEX.KG, a derivative of pGEX-2T (Guan and 
Dixon, 1991; Lakkis et al., 1996). Lakkis et al. (1996) determined from N-terminal 
sequencing that the expressed protein from the glutathione S-transferase expression system 
contains two extra amino acids at the amino terminal end from the thrombin cleavage site. 
The resulting plasmid, pGEXmCA7, was sent to us from Dr. Tashian's laboratory. 
pGEXmCA7 was transformed into Escherichia coli (DH5a) and rMCA VII protein 
expressed and purified as follows (Smith and Johnson, 1988; Guan and Dixon, 1991; Lakkis 
et al., 1996). IPTG was added to a final concentration of 0.4 mM to induce expression of 
rMCA VII in DH5a cells grown in 2 x YT medium (with ampicillin). Frozen DH5a cells 
containing the expressed rMCA VII were thawed in a solution of PBST containing 2 mM 



Abbreviations: rMCA VII, recombinant murine carbonic anhydrase VII; MCA Vllb, a 
truncated form of murine carbonic anhydrase VU lacking 23 residues from the amino-terminal 
end; Mes, 2-(N-morpholino) ethanesulfonic acid; Mops, 3-(N-morpholino)propanesulfonic 
acid; Hepes, N-[2-hydroxyethyl]piperazine-N'-2-ethanesulfonic acid; Taps, 3- 
[tris(hydroxymethyl)methyl] aminopropanesulfonic acid; Ches, 2-(cyclohexylamino) 
ethanesulfonic acid; IPTG, isopropyl-P-D-thiogalactoside; PBST is the solution containing 
150 mM NaCl, 16 mM Na 2 HP0 4 , 4 mM NaH 2 P0 4 , pH 7.3 and 1% Triton X-100; SHffi, 
solvent hydrogen isotope effect. 





o 


ID 






O <XI 


en 


CM 






00 Irt 


E-. CJ CD 
CD CJ 


Q 






O CD 

o 3 

E» < CO 


CD U 








CD CJ 
b < 


CD U 


CD 






< E- 2 


CD U 








CJ CD 


< H 








CD O 


U CD 


O 




t 

XI 

1—4 


E- < U 


U CD 
U CD 






CJ CD 
U CD 


CD O < 




CD U < 


z$ 






> 


CD CJ 

< E- 


< E- 


M 




CD O U 


O CD 






< 


H < 


O CD 






< E- 


O CD 


a, 


X 


z 


E- < >- 


E- < 






O CD 


< E- 








CJ E- < 


E- < >■ 






E- 2 lu 


CD CJ 








r «s 


E- < 








E- < 


O CD 


►J 






O CD J 


CD O 








°iP 


it 








x. 






CD U U 


O CD 








CD CJ 


< E- 








P < 


CJ CD 


X 






CJ CD J 


CD U 








< E- 


CD CJ 








CJ CD 


£2 


2 






CJ CD Du 








< CD CJ 


"? E- 








< E- 


O < CD 


2 


X 




CJ CD O" 


CD « CJ 








CD CJ 


U CD 








H < 


uf-2 


co 






O CD J 








CJ CD 


O CD 








CD CJ 


O CD 


a, 






< E-> CO 


U CD 








CJ CD 


CD CJ 








CJ CD 


CD CJ 


o 






CJ CD 0-> 


O CD 








E- CD CJ 


< E- 








CJ CD 


CD O 









E- < CO 


CJ CD 








CJ CD 


< E- 








< E- 


CD CJ 


Q 






E- < » 


CD CJ 








CD CJ 


< E- 








E- < 


CJ CD 


a 






CD CJ > 


u B 








E- < 


CD CJ 








CJ CD 


CD CJ 


C3 






CD U < 


CJ CD 








CD CJ 


< 










< E- 


t- 




i^. 


>- 




o cd a 


u 










CJ CD 


e> 










CD CJ 


o 




r. 


CD 




< E-> CO 


a 










CJ CD 


a 










CJ CD 


H 




c^ 


S 




E- < CO 


U 










CJ < E- 


Q 










E- < 


o 




c- 


o 




< E- w 


u 










O CD 


< 










E- < 


u 




c^. 


X 




< E- M 


u 










E- < 


< 
u 




c^. 


X 




< E- 2 


u 










CJ CD 













H < 


o 




f^. 


CD 




< E- M 


u 










CJ CD 


u 








t 


CJ CD 


< 




r^ 


H 


CJ CD Qj 


o 










< E- 


E- 








t-H 


CJ CD 


< 




c^. 


E 


M 


E- < CO 


Q 






> 


< CD U 
< E- 


o 




CO 


< 

u 


O CD a 
CJ CD 
CD CJ 


CD 






CD 


ki 


CJ CD DC 



o 

CSI 



U 



CD 



CD CJ 
CD CJ 
CD CJ 

It 

CD CJ 

CD U 

E- < 

CJ CD 

CJ CD 

CJ CD 

CJ CD 

CJ CD 

CD CJ 

CD CJ 

CD CJ 

CD CJ 

CD U 

E- < 

U CD 

H < 

CD CJ 

E- < 

CD CJ 

CD CJ 

E- < 

CD CJ 

CJ CD 

CJ CD 

it 

CD CJ 

< E- 
CJ CD 

< E- 
CD CJ 

t < 

< E- 
CD CJ 

1 E- < 

CD CJ 

< E- 
CJ CD 

< E- 
CD O 
H < 

sr: 

CJ CD 

El 

CJ CD 

< H 
CD O 
CD CJ 
H < 
CD O 
CD CJ 

< E- 
U CD 
U CD 
E- < 
CD CJ 
E- < 



O CD 
CD CJ 
CD CJ 

t: < 

St 

O CD 

$£ 

V CD 

CJ CD 

< E- 
O CD 
E- < 

< E- 
CJ CD 
CD CJ 

< E- 
CJ CD 
E- < 
U CD 



CD 
U 

Ou 

CD 

CD 

CO H 

> 

> 

E- 

te 

a 
a 

CO 

a 

2 
b 

Q 
> 

> 
to 

as 

CD 

2 
2 
E-> 

M 

CO 

■J 



O VO 

en -4 



^8 

< E- 

CJ CD 

CJ CD 

U CD 

CJ CD 

££ 

CJ CD 

CJ CD 

E* < 

CD CJ 



CD 



CO 



CO 



CD 



CD U 

< E- 
U CD 

CJ CD 

< E- 
O CD 
CD CJ 

< E- 



u 



CO 



CD 



CJ 



U 
E- 
CD 
< 
CD 
□ 
U 

u 
< 

CD 

„ b 

CJ CD 

CD CJ 

CJ CD 

CD CJ 

It 

CD O 

sr: 

CJ CD 

CD CJ 

CD CJ CD 

CD CJ 

CD CJ 

E- < S 

CJ CD 

< H 

CJ CD X 
CJ CD 

E- 2 U. 
CJ CD 

< E- 

CJ CD X 
CJ CD 

«t 

CD 

CJ 

E- 

CD O 
CD CJ 

< h k: 

CJ CD 

CD J 

CJ CD 

CD CJ 

O CD OS 

t < 

< E- 

E- < >- 
O CD 
" CD 
(D Cu 



r > 



CX X 



E- 
O 
CD 
< 
U 



j u, 



o 

IT) 



CD 



U 



Cu 



i— i 3 C C 

on -a .5 .5 

r - B s 



e3 



8 



0) 



U 



21 



S x c 

O 4> B 

8 £ a 



S3 t3 

js aj 



o 

U 



if) 



CM 



«s 



o 



<U 



22 



O ID 

v r- 



O VO 
el o 
V© CM 



o «o 

CM CI 

f CM 



o o 



O ex 




D 

c 
o 
o 



v r- 






O 
cs: 



CI 
CM 



tN 

V 

3 
60 



23 
EDTA, 0.2 mM phenylmethylsulfonyl fluoride, 5 mM benzamidine, 0.4 mM MgCl 2 , 0.4 mM 
ZnS0 4 , 0.1% p-mercaptoethanol, 0.1 mg/mL deoxyribonuclease I and 0.5 mg/mL lysozyme 
with stirring for two hours at 4°C. After cell lysis the cell debris was pelleted by 
centrifugation at 23,000 x g for 1 hour at 4°C. The supernatant was subjected to two affinity 
chromatography steps. First, the supernatant containing the glutathione S-transferase/rMCA 
VII fusion protein was stirred with 10 mL swollen glutathione S-agarose beads (Sigma 
Chemical Company) and then equilibrated in PBST at 4°C to allow binding of the fusion 
protein to the beads. The beads were washed with cold PBST and 20 mL of thrombin 
cleavage buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 2.5 mM CaCl 2 , and 0.1% p- 
mercaptoethanol) was added to equilibrate the beads for thrombin cleavage. Thrombin 
cleavage of the fusion protein was achieved by the addition of 10 U of thrombin in thrombin 
cleavage buffer and incubation at room temperature for 30 min. rMCA VII was collected in 
the eluate and dialyzed against 1 mM Tris (pH 8.0). The second purification step involved 
affinity chromatography on a gel containing p-aminomethylbenzenesulfonamide coupled to 
agarose beads according to Khalifah et al. (1977) with minor modifications. The enzyme was 
stored at 4 °C, -20 °C or -70 °C for several months with 100% recovery of the original 
activity that was determined during its purification at 4 °C. 
Subcloning and Site-Directed Mutagenesis of CA VII 

P23M truncation and P23M/H64A truncation mutants of CA VII. A truncated CA 
VII and a truncated H64A CA VII mutant were prepared by Dr. Minzhang Qian in the 
laboratory of Dr. Philip J. Laipis (University of Florida) and their methods for producing these 
mutants are as follows. Oligonucleotides which introduced 1) an Nde I restriction site 



24 
containing a methionine codon at position 998 in pGEXmCA7, changing the Pro (CCC) 
codon at the cDNA amino acid position 23 to a Met codon (DS2 1 5 : CAC AAAGCTGC ATA 
TGATTGCCCAGGG), and 2) a Bel I restriction site (including the natural stop codon) at 
position 1728 in pGEXmCA7 (ATGGAGTCTTGATCAGGCCTGGAA) were synthesized 
and used in four separate PCR reactions to amplify the truncated form of murine CA VII 
(designated MCA Vllb throughout this chapter). This truncation removes the first 23 amino 
acids of the original rMCA VII cDNA. The products of the PCR reactions were cloned into 
pGEM-T (Promega) and four independent clones from the separate PCR reactions sequenced. 
Two clones (all four contained the correct sequence) were then mutated using single-stranded 
DNA template and a mutating oligonucleotide (AGAGGATTCACATGGTGGACA) to 
remove the naturally occurring Nde I site at position 724 in the rMCA VII cDNA insert 
(Kunkel et al., 1987). The mutated inserts were removed from pGEM-T by digestion with 
Nde I and Bel I and inserted into a Nde I and BamH I cleaved pET3 1+ expression vector 
(Tanhauser et al., 1992) to allow efficient high-level expression and site-directed mutagenesis. 
This expression plasmid constructed for the truncation mutant is termed pET31+MCA7b. 
Construction of the His 64 to Ala mutant form of MCA Vllb, the truncated mutant, was as 
previously described using a mutating oligonucleotide and a uracil-containing single-strand 
template (Kunkel et al., 1987; Tanhauser et al., 1992). This expression plasmid constructed 
for the truncation mutant containing the His 64 to Ala mutation is termed pET3 l+MCA7b 
H64A. 

H64A mutant of CA VII . A full-length wild-type CA VII that is lacking the two extra 
amino acids at the amino-terminal end of the protein that are normally produced by the 






25 
glutathione S-transferase expression system and the H64A mutant of full-length CA VII were 
prepared and placed in a pET3 1+ expression vector by Ms. Nina Wadhwa in the laboratory 
of Dr. Philip J. Laipis (University of Florida) and their methods for producing these mutants 
are as follows. An oligonucleotide (DS225: CGCGTGGACATATGACCGGCCATCAC) 
which introduced an Nde I restriction site containing a methionine codon at position 932 (the 
normal start position in the absence of the two thrombin cleavage residues) in pGEXMCA7 
was synthesized and used with DS214 in separate PCR reactions to amplify the full-length 
form of rMCA VII. The products of the PCR reactions were cloned into pGEM-T (Promega) 
and individual clones from separate PCR reactions sequenced. Three individual clones with 
correct sequence were digested with Nde I and Pvu II, releasing a 471 bp fragment containing 
the first 157 amino acids of rMCA7. This fragment was cloned into a similarly digested 
pET3 l+MCA7b expression plasmid. This effectively replaced the truncated amino terminus 
of the MC A7b insert with the full length sequence while eliminating the necessity for both 
site-directed mutagenesis to remove the Nde I and Bam HI sites and sequencing of more than 
the first 471 nucleotides. This expression plasmid constructed for the full-length CA VII is 
termed pET31+MCA7. The His 64 to Ala mutant of the full-length rMCA VII (without the 
two thrombin cleavage residues) in the pET3 1+ system was derived from pET3 l+MCA7b 
H64A by removing a 274 bp Pst I (bp 638-912) fragment and inserting it into a 
dephosphorylated, Pst I digested pET31+MCA7 full-length backbone from which the 
corresponding fragment (704-978) had been removed. 

The kinetic properties of rMCA VII expressed from the pET3 1+ expression vector, 
which lacks the two addition amino-terminal thrombin cleavage residues, was unchanged from 









26 
that of rMCA VII, which contains two extra amino acids at its amino-terminal end. 
Therefore, unless otherwise indicated rMCA VII protein expressed from the glutathione S- 
transferase expression system was used for the experiments in this chapter. 

Expression and purification of CA VII proteins expressed from the pET31+ vector. 
The various forms of rMCA VII and MCA Vllb were expressed in E. coli strain 
BL21(DE3)pLysS as described (Tanhauser et al., 1992). DNA from the expression plasmids 
used to produce each protein was sequenced to confirm the structure of each insert. 
Purification was performed through previously described procedures using affinity 
chromatography on a gel containing /;-aminomethylbenzenesulfonamide coupled to agarose 
beads (Khalifah et al., 1977; Heck et al., 1994). The enzymes were then stored at 4°C. 
Enzyme Purity 

Electrophoresis on a 10% polyacrylamide gel stained with Coomassie Blue was used 
to confirm the purity of all of the CA VII enzyme samples. All enzyme samples used in the 
kinetic experiments were greater than 95% pure. Active CA VII enzyme concentration was 
determined by inhibitor titration of the active site with ethoxzolamide (Kj = 0.5 nM) by 
measuring I8 exchange between CO, and water (see below). The molar absorptivity at 280 
nm was determined to be 2.6 x 10 4 M" 1 cm" 1 for rMCA VII. 
Stopped-Flow Spectrophotometry 

Initial velocities were determined by following the change in absorbance of a pH 
indicator (Khalifah, 1971) at 25 °C using a stopped-flow spectrophotometer (Applied 
Photophysics Model SF. 17MV). C0 2 solutions were made by bubbling carbon dioxide into 
water or D 2 for the solvent hydrogen isotope effect studies. The maximum concentration 



27 
of C0 2 in H 2 achieved by this method was 17 mM following dilution (Pocker and 
Bjorkquist, 1976). Dilutions were made through two syringes with a gas tight connection, 
the CO, concentrations ranged from 1.7 to 17 mM in H 2 0. For the dehydration direction, 
KHCOj was dissolved in degassed water and the HCO," substrate concentrations ranged from 
2.8 to 25 mM in H 2 0. Final buffer concentrations were 25 mM and Na 2 S0 4 was used to 
achieve a final ionic strength of 0.2 M. The buffer-indicator pairs, pK a values and the 
wavelengths observed were as follows: Mes (pK, = 6. 1) and chlorophenol red (pK, ■ 6.3) 574 
nm; Mops (pK, = 7.2) and/Miitro phenol (pK a = 7. 1) 401 nm; Hepes (pK a = 7.5) and phenol 
red (pK, = 7.5) 557 nm; Taps (pK, = 8.4) and m-cresol purple (pK, ■ 8.3) 578 nm; Ches (pK, 
= 9.3) and thymol blue (pK a = 8.9) 596 nm. For each substrate at each pH, the mean initial 
velocity was determined with at least 6 traces of the initial 5-10% of the reaction. The 
uncatalyzed rates were determined in a similar manner and subtracted from the total observed 
rates. The kinetic constants, k cat and k^/K,,,, and the apparent ionization constants from their 
pH profiles were determined by a nonlinear least-squares methods using Enzfitter (Elsevier- 
Biosoft). Standard errors in k^and k^/F^ were generally in the range of 5% to 20% and 1% 
to 10% respectively. 
lg O Exchange 

The rate of exchange of 18 between species of C0 2 and water (equations 2-1 and 2- 
2) is catalyzed by carbonic anhydrase (Silverman, 1982). 

HC00 18 0- + EZnH 2 ** EZnHC00 18 0" + H 2 ** COO + EZn 18 OH- + H 2 (2-1) 

k B + H 2 

EZn 18 OH" + BH + ** H + EZn 18 OFT + B * EZnH 2 18 ** EZnH 2 + H 2 18 (2-2) 



28 
An Extrel EXM-200 mass spectrometer with a membrane permeable to gases was used to 
measure the exchanges of 8 shown in equations 2-1 and 2-2 at chemical equilibrium and 25 
°C (Silverman, 1982). Solutions contained no added buffers and total ionic strength was 
maintained at 0.2 M with Na 2 S0 4 unless otherwise indicated. 

This method determines two rates in the catalytic pathway (Silverman, 1982). The first 
is R, the rate of interconversion of C0 2 and HC0 3 ~ at chemical equilibrium. Equation 2-3 
shows the substrate dependence of R,. 

R,/[E] - k cat "[S]/(K eff s + [S]) (2-3) 

Here [E] is the total enzyme concentration, k^ 6 * is a rate constant for maximal HC0 3 " to C0 2 
interconversion, [S] is the substrate concentration of HC0 3 " and/or C0 2 , and K eff s is an 
apparent substrate binding constant (Simonsson et al., 1979). Equation 2-3 can be used to 
determine the values of k cat e 7K eff s when applied to the data for varying substrate 
concentration, or to determine k cal e YK eff s directly from R,/[E] when [S] « K eff s . In the studies 
reported here, the values of R,/[E] as a function of total concentration of all species of C0 2 
was linear at ([C0 2 ] + [HC0 3 - ]) as large as 200 mM. This indicates that [S] « K,. ff s and under 
this condition k^/K^ can be obtained directly from R,/[E]. Under steady- state conditions, 
when [S] « K,,, all enzyme species are at their equilibrium concentrations. Hence in both 
theory and practice, k^VK,/ 02 is equivalent to k^JK^ for C0 2 hydration as measured by 
steady-state methods (Silverman, 1982; Simonsson et al., 1979). The kinetic constant k^/I^ 
and the determination of apparent ionization constants from the pH profile were carried out 
by nonlinear least-squares methods using Enzfitter (Elsevier-Biosoft). 

This method also determines a second rate which is the rate of release from the 
enzyme of water labeled with 18 designated R^q (equation 2-2). A proton from a donor 



29 
group BH converts the zinc-bound 18 0-labeled hydroxide to zinc-bound H 2 18 , which readily 
exchanges with unlabeled water and is greatly diluted into the solvent water H 2 16 0. The value 
of Rhjo can be interpreted in terms of the rate constant from a predominant donor group to 
the zinc-bound hydroxide according to equation 2-4 (Silverman et al., 1993), in which k B is 
the rate constant for proton transfer to the zinc-bound hydroxide, K B is the ionization constant 
for the donor group and K E is the ionization constant of the zinc-bound water molecule. For 
our data Rh2c/[E] was determined by equation 2-4 using the program Enzfitter (Elsevier- 
Biosoft). 

R H2 o/[E] = k B /{(l + K B /[H + ])(1 + [FT]/K E )} (2-4) 

For solvent hydrogen isotope effects, R H2 c/[E] an d k^/K,,, were determined in 99.8% 
D 2 0. All pH measurements are uncorrected pH meter readings. This is based on the 
assumption that the correction of a pH meter reading in 1 00% D 2 to obtain pD (pD = pH 
+ 0.4) is approximately canceled by the change of pK a in D 2 for almost all acids with pK a 
between 3 and 10. 
Inhibition 

Inhibition by ethoxzolamide and acetazolamide was determined using 18 exchange 
at chemical equilibrium. The values of K< in Table 2-1 were obtained by least-squares fit of 
catalytic velocity to the expression for competitive inhibition as a function of inhibitor 
concentration under the conditions that the total substrate concentration ([C0 2 ] + [HC0 3 "] 
= 25 mM) was much less than the apparent binding constant for total substrate, K eff s . At the 
pH of these measurements, 7.3 to 7.5, K eff s is greater than 100 mM for CA VII. The values 
of Kj determined from R H2C /[E] ' n the same manner agreed to within 25% with the values 
determined from R,/[E]. 



30 



Table 2-1: Inhibition Constants K< (Nanomolar) for Isozymes of Carbonic Anhydrase 
Determined by 18 Exchange. 



Isozyme 



K, (nM) 



Acetazolamide 



Ethoxzolamide 



Reference 



Human CA II 



Human CA III 



Murine CA IV 



Murine CA V 



Murine CA VII 



60 



40.000 



60 



16 



8 


LoGrasso etal, (1991) 


too 


LoGrasso etal, (1991) 


16 


Hurt etal, (1997) 


5 


Heck etal, (1994) 


0.5 


This work 



Note: All values of Kj were determined at pH 7.2-7.5 and 25 °C. 






31 
Hydrolysis of 4-Nitrophenyl Acetate 

Measurement of the esterase function of rMCA VII was performed by the method of 
Verpoorte et al. (1967) by following the absorbance change at 348 nm, the isosbestic point 
of 4-nitrophenol and its conjugate base, nitrophenolate ion. Concentrations of buffer and 
Na 2 S0 4 were both 33 mM and initial velocities were determined at 25 °C. The uncatalyzed 
rates were subtracted from the observed rates, and the kinetic constant k^/K,,, and the 
determination of apparent ionization constants from the pH profile were carried out by 
nonlinear least-squares methods using Enzfitter (Elsevier-Biosoft). 

Results 

Recombinant Murine CA VII 

The full length recombinant murine CA VII protein (rMCA VII) used in this study has 
a portion of its amino-terminal amino acid sequence derived from the human CA VII cDNA 
(Lakkis et al., 1996). More specifically, Ling et al. (1995) obtained a 91% complete mouse 
CA VII cDNA by RT-PCR using RNA isolated from adult mouse (C57/BL6) brain. This 
cDNA lacked sequence encoding the 28 amino-terminal residues. Lakkis et al. (1996) then 
constructed an almost complete murine CA VII cDNA by carrying out PCR under relatively 
stringent conditions with a 5' primer from the human CA VII sequence starting at the initial 
ATG and extending 26 nucleotides into the gene (Lakkis et al., 1996), and a 3' primer 
determined by the murine CA VII cDNA of Ling et al. (1995) (see Figure 2-1 for location of 
5' and 3' primers). This construct allowed the determination of an additional 58 murine 
nucleotides after the 26 human 5' primer nucleotides at the 5' end of the murine CA VII 
cDNA (Figure 2-1). This extended murine CA VII cDNA sequence showed that 18 of the 19 



32 
newly-derived amino acids are identical between human and murine CA VII (Figure 2-1). 
However, the rMCA VII cDNA used in this study has 26 nucleotides (including the start 
ATG) that are of human origin (Figure 2-1) and thus encodes a chimeric CA VII protein with 
the 9 N-terminal amino acids of human origin and the remaining 253 from the mouse. The 9 
amino-terminal residues of the murine CA VII protein remain unknown; however, in view of 
the highly conserved nature of the human and murine CA VII sequences, it is likely that they 
will be identical to the human sequence. 
Catalytic Activity 

The ratio k^/K,,, for rMCA VII determined from 18 exchange between C0 2 and 
water yielded a pH profile that was best fit by the sum of two ionizations (Figure 2-2, Table 
2-2; Tipton and Dixon, 1979). A very similar result to these was obtained by stopped-flow 
spectrophotometry in which the pH profile of kJK„ of rMCA VII for C0 2 hydration over 
the pH range of 5.3 to 9.0 was also described by two ionizations with values of pK a of 6.2 ± 
0.9 and 7.6 ± 0.5 and with a maximum of (3.3 ± 1.0) x 10 7 M" 1 s" 1 at high pH (Data not 
shown). The maximal values of k c JK ni for hydration measured by stopped-flow 
spectrophotometry were somewhat lower than those measured by 18 exchange. The results 
of the steady-state measurements of k ca /K m for rMCA VII in the HC0 3 " dehydration direction 
yielded a maximum at low pH and was dependent on a single ionizable group (Figure 2-2, 
Table 2-2). Because of the unfavorable equilibrium between C0 2 and HC0 3 ', the 
measurements in the dehydration direction were not extended to regions above pH 7.2. 
Hence, in the dehydration direction the second ionization at pH near 7.5 was not observed. 

The pH dependence of k^/K^ determined from 18 exchange between C0 2 and 
water, was also measured for the full-length rMCA VII H64A mutant. The pH profile of 



33 



</> 







Figure 2-2. The pH dependence of k^/K,,, for (•) hydration of C0 2 and (O) dehydration of 
HC0 3 " catalyzed by rMCA VII at 25 °C. KJi^ for hydration was obtained by 18 exchange 
using solutions containing no buffers and in which the total ionic strength of solution was 
maintained at 0.2 M by addition of NajSO* and the total concentration of all species of C0 2 
was 25 mM. A nonlinear least-squares fit to the data points for C0 2 hydration is represented 
by the solid line. The fit was to two ionizations with values of pK a and maximal k^/F^ given 
in Table 2-2. The dashed line is a nonlinear least-squares fit to one ionization with a pK, = 7. 1 
± 0. 1 and a maximal value of KJKn at (7.2 ± 0.3) x 10 7 M' 1 s 1 . The ratio k cat /K m for 
dehydration of HC0 3 " was obtained by stopped-flow spectrophotometry in the presence of 
25 mM of one of the following buffers: pH 5.3-6.4, Mes; pH 6.6-7.2, Mops; pH 6.9-7.2, 
Hepes. Total ionic strength of solution was maintained at 0.2 M with Na^O,,. The solid line 
is a nonlinear least-squares fit of the data points to a single ionization with pK a and maximal 
k^/K,,, given in Table 2-2. 



34 



Table 2-2: Catalysis by Carbonic Anhydrase VII: Maximal Values of Steady-State Constants 
with Values of Apparent pK a . 



kJKn (NTs' 1 ) pK a(kcatKm) k^, (s" 1 ) pK 



a(kcat) 



Hydration of CO, (7.6 ±0.3) x 10 7 6.2 ±0.5 (9.4 ± 2.4) x 10 5 ~8.2 a 

7.5 ±0.3 6.2 ±0.2 



Sc 



Hydration of CO, (8.2 ± 0.3) x 10 7 7.7 ±0.1 (4.5 ± 0.4) x 10 

(MCA VIIb) b 

Hydration of C0 2 (9.7 ± 0.2) x 10 7 7.2 ±0.2 (1.6 ± 0.2) x 10 5 8.9 ±0.2 

(H64A) d 

Hydrolysis of 71±8 5.3 ± 0.3 

4-Nitrophenyl Acetate 7.1 ± 0. 1 

Dehydration of HC0 3 " (9.7 ± 1.0) x 10 6 6.8 ±0.2 (1.9 ±0.1) x 10 5 6.7 ±0.1 

Note: All data was obtained using rMCA VII protein except where indicated. Experimental 

conditions as given in the corresponding Figure legends. 

a Because of the uncertainty in the maximal value of k ca „ this pK a is poorly determined. 

b This truncated form of murine CA VII is lacking the amino-terminal 23 residues of rMCA 

VII shown in Figure 2-1. 

c Not measured. The maximal value of Ic^ was determined from one measurement only, at pH 

9.1. Therefore, a value of pK a(kcat) could not be determined. 

d This is the H64A mutant of full-length rMCA VII. 



35 
k^/Kn, was described by one ionization (Table 2-2). A very similar result to these was 
obtained by stopped-flow spectrophotometry in which the pH profile of k^/Kn, for C0 2 
hydration over the pH range of 5.9 to 9.5 was also described by one ionization with a value 
of pK a of 7.6 ± 0.6 and with a maximum of (2.4 ± 0.6) x 1 7 M" 1 s" 1 at high pH (Data not 
shown). The maximal value of k^/K,,, for hydration measured by stopped-flow 
spectrophotometry were somewhat lower than those measured by 18 exchange. 

The pH dependence of k^/K^ for two truncated forms of murine CA VII were also 
measured from 18 exchange between C0 2 and water. One truncated form had the amino- 
terminal 23 residues removed, with a new amino-terminus starting at the Pro 23 — » Met 
mutation; it is designated MCA Vllb (Figure 2-1). MCA Vllb was further mutated by 
replacing His 64 with Ala. MCA Vllb had values of k^/K,,, for C0 2 hydration adequately 
fit to a single pK a with values given in Table 2-2. The pH profile of k^/K,,, for MCA Vllb 
H64A was identical. 

Measurements of C0 2 hydration by stopped-flow spectrophotometry gave a maximum 
value of k^ at pH > 9 for full-length rMCA VII and the H64A mutant (Figure 2-3, Table 2- 
2). The data for k cal over the pH range of 5.3 to 9 was best fit to two ionizations for wild- 
type rMCA VII whereas the rMCA VII H64A mutant had a pH profile of k ca( described by 
one ionization (Figure 2-3, Table 2-2). Steady-state measurements for k cat for dehydration 
of HC0 3 " showed a maximum at low pH for wild-type rMCA VII (Figure 2-3, Table 2-2). 

Values of the 18 0-exchange parameter R H 2c/[E] describe the proton-transfer 
dependent rate of exchange of H 2 18 into solvent water (equation 2-2; Silverman, 1982). As 
was found for human CA II (Silverman et al., 1993), for rMCA VII the pH profile of R^E] 



36 




Figure 2-3. The turnover number k cat for (•) hydration of CO, and (O) dehydration of 
HC0 3 " catalyzed by rMCA VII and (A) hydration of C0 2 catalyzed by rMCA VII H64A 
obtained by stopped-flow spectrophotometry at 25 °C in the presence of 25 mM of one of 
the following buffers: pH 5.3-6.4, Mes; pH 6.6-7.2, Mops; pH 6.9-7.5, Hepes; pH 7.7-8.3, 
Taps; pH 8.6-9.1, Ches. Total ionic strength of solution was maintained at 0.2 M with 
Na 2 S0 4 . The solid line is a nonlinear least-squares fit to two ionizations for the data points 
in C0 2 hydration catalyzed by rMCA VII and for rMCA VII H64A the solid line is a nonlinear 
least-squares fit to one ionization with the value of pK a and maximal k^/Isn f° r both enzymes 
given in Table 2-2. The solid line through the data points for the dehydration of HC0 3 " is a 
nonlinear least-squares fit one ionization with pK a values and maximal k^, also given in Table 
2-2. 



37 
was bell-shaped with a maximum occurring near pH 7 (Figure 2-4). The H64A mutant of 
rMCA VII was found to have much reduced values of RwTE] at pH < 8 as compared to 
rMCA VII (Figure 2-4). The inset of Figure 2-4 shows the pH profile for the differences in 
Rjco/tE] between full-length rMCA VII and rMCA VII H64A. The shape and magnitude of 
this difference plot reflects the loss of a proton donor or donors of pK a near 7. 1 . 

R^o/fTE] for rMCA VII was inhibited by CuS0 4 with an inhibition constant of 0.33 
uM at pH 7.5 and 25 °C; the addition of CuS0 4 , up to a final concentration of 40 uM, to 
rMCA VII had no effect on R,/[E] (Data not shown). 

The truncation mutants of CA VII had much reduced values of R H2C /[E] at pH < 8 
(Figure 2-5). The values of Rh2c/[ e ] catalyzed by truncated MCA Vllb H64A were reduced 
even further in this region of pH (Figure 2-5). It must be noted that in Figure 2-5 the data 
of Rh2c/[E] for MCA V 1115 H64A was not adequately fit to a single ionization at pH less than 
7. This deviation of the data from a single ionization behavior may reflect enzyme 
denaturation at low pH or perhaps the removal of the amino terminus and histidine 64 has 
generated a mutant with a more complex pH profile of Rfec/tE] at low P H If tnis is tne case ' 
then this poor fit is a result of the lack of consideration of other influences or ionizations that 
may contribute to the pH dependence of R H20 /[E] for this mutant. In either case, the inset of 
Figure 2-5 shows the pH profile for the differences in R H2C /[ E ] between the full-length rMCA 
VII and the truncated MCA Vllb, and between MCA Vllb and MCA Vllb H64A. The shape 
of these two difference plots is very similar reflecting in each case the loss of a proton donor 
or donors of pK a near 6.9 to 7.5. 



38 



10 



Difference Rh2o/[E] 

(x 10 s s 1 ) 2 h 

1 




pH 



Figure 2-4. Variation with pH of RwTE], tne proton-transfer dependent rate constant for 
the release from the enzyme of 18 0-labeled water, catalyzed by (•) rMCA VII and (A) the 
H64A mutant of rMCA VII. Solutions contained no buffers and the total ionic strength of 
solution was maintained at 0.2 M by addition of NajSO,,; the total concentration of all species 
of C0 2 was 25 mM. The solid line is a fit of equation 2-4 to the data for rMCA VII with 
values ofpK, for the proton donor of 7. 4 ±0.1 and acceptor groups of 5. 8 ±0.1 and 1^, the 
rate constant for proton transfer to the zinc-bound hydroxide in the dehydration direction of 
catalysis, of (3.2 ± 0.4) x 10 5 s' 1 . The solid line representing the fit of equation 2-4 to the data 
for rMCA VII H64A yielded pK a values for the proton donor of > 9 and acceptor groups of 
6.5 ± 0. 1 and ke = (2.6 ± 0. 1) x 10 4 s '. Inset. The difference in R^o^E] between rMCA VII 
and rMCA VII H64A. The solid line is a nonlinear least-squares fit describing proton transfer 
(equation 2-4) from a donor group of pK a 7.1 ±0.1 and zinc-bound hydroxide the conjugate 
acid of which has pK, 5.9 ± 0. 1 with a rate constant for proton transfer k B = (3.3 ± 0.3) x 10 5 
s 1 . Note, at pH > 8.5 the fit of the data for wild-type and H64A cross. This is a result of the 
fit of rMCA VII to only two ionizations and our lack of consideration of the addition 
contribution of a proton donor group at pH > 8.5, that most certainly exists in the wild-type. 



39 




Figure 2-5. The variation with pH of K mo /[E] catalyzed by (•) full-length rMCA VII, (A) 
the H64A mutant of full-length rMCA VII, (□) the truncated MCA Vllb (see Figure 2-1), 
and (■) a H64A mutant of the truncated MCA Vllb. Experimental conditions are as 
described in Figure 2-4. Inset: (□) The difference in Rh2o/[ e ] between rMCA VII and the 
truncated form MCA Vllb. The solid line is a nonlinear least-squares fit describing proton 
transfer (equation 2-4) from a donor group of pK a 6.9 ± 0.2 and zinc-bound hydroxide the 
conjugate acid of which has pK, 5.9 ± 0.2 with a rate constant for proton transfer of 1^ = (3.4 
± 0.8) x 10 5 s 1 . (■) The difference in R H2 o / [ E ] between MCA Vllb and the H64A mutant of 
MCA VTIb. The solid line is a nonlinear least-squares fit describing proton transfer (equation 
2-4) from a donor group of pK a 7.5 ± 0.2 and zinc-bound hydroxide the conjugate acid of 
which has pK, 6.2 ± 0.2 with a rate constant for proton transfer of k B = (5.5 ± 1 . 1) x 10 4 s" 1 . 



40 
rMCA VII is able to catalyze the hydrolysis of 4-nitrophenyl acetate. The pH profile 
of kc^/Kn, has a maximum at pH > 9 and is best fit to two ionizations (Figure 2-6, Table 2-2) 
similar to k^/K^ in the C0 2 hydration direction. In some isozymes of CA there is a 
nonspecific esterase activity not associated with the zinc, as has been found in CA I (Wells 
et al., 1975) and CA III (Tu et al., 1986). However, that is not the case for rMCA VII; at a 
variety of different pH values the esterase activity was found to be greater than 98% inhibited 
in the presence of equimolar concentrations of the active-site inhibitor ethoxzolamide (1^ = 
0.5 nm) and enzyme (present at 10~ 7 M). 

Inhibition of l8 exchange between C0 2 and water catalyzed by rMCA VII with two 
classic sulfonamides acetazolamide and ethoxzolamide was tested. The resulting values of the 
inhibition constant ¥^ are compared to the inhibition values of other isozymes of CA in Table 
2-1. 

The solvent hydrogen isotope effects (SHIE) observed for catalysis by rMCA VII 
were 1 .0 ± 0. 1 for kJK„, for CO, hydration at pH 6.8. On k^, the SHIE was 3 .0 ± 0. 1 at pH 
6.8 in solutions containing 25 mM Hepes consistent with rate-determining proton transfer 
involving the aqueous ligand of the zinc (equation 1-2, Chapter 1, page 9). The SHIE at pH 
6.8onR H2O /[E]was3.3±0.4. 

Discussion 

Comparison of Isozymes of Carbonic Anhydrase 

I have compared the steady-state catalytic constants of CA VII with CA II, the most 
efficient of the carbonic anhydrase isozymes, and with five other isozymes in the a-class of 



41 




PH 






Figure 2-6. The pH dependence of k M /K m for the hydrolysis of 4-nitrophenyl acetate 
catalyzed by rMCA VII. Data were obtained at 25 °C in 33 mM of one of the following 
buffers: pH 5.3-6.5, Mes; pH 6.9-7.2, Mops; pH 7.3-7.7, Hepes; pH 8.1-8.9, Taps; pH 9.1- 
9.4, Ches. The solid line is a nonlinear least-squares fit of two ionizations with pK a values and 
maximal k^/K^ given in Table 2-2. 



42 
the carbonic anhydrases. The steady-state constant k^/K,,, contains the rate constants up to 
and including the first irreversible step, which is the departure of HC0 3 ~; these are the steps 
of equation 1-1 (Chapter 1, page 9). The pH dependence of k^/K,,, describes the ionization 
state of the zinc-bound water (Christianson and Fierke, 1996; Lindskog, 1997). In the C0 2 
hydration direction, the maximal value of KJKn of 7.6 x 10 7 M" 1 s" 1 for rMCA VII is half that 
of CA II but somewhat greater than those for CA I, CA IV, and CA V (Table 1-2, Chapter 
1, page 10). The observed solvent hydrogen isotope effect of 1.0 ± 0. 1 on k^/K^, indicates 
no rate-contributing proton transfer in the steps of equation 1-1 (Chapter 1, page 9) and is 
consistent with a direct nucleophilic attack of the zinc-bound hydroxide on C0 2 (Lindskog, 
1997); in this respect rMCA VII is similar to CA II and the other isozymes in the cc-class. 

The maximal value of the turnover number k cal for hydration of C0 2 catalyzed by 
rMCA VII approaches that of CA II (Table 1-2, Chapter 1, page 10). The value of k ca , 
contains rate constants for the steps from the enzyme-substrate complex through the proton 
transfers of equation 1-2 (Chapter 1, page 9). The kinetic constants for rMCA VII place it 
among the most efficient of the carbonic anhydrases with 67% of the activity of CA II. 
Considering the similarity in steady-state constants for C0 2 hydration catalyzed by rMCA VII 
and C A II, it is interesting that the capacity of rMCA VII to catalyze the hydrolysis of 4- 
nitrophenyl acetate (maximal KJK^ = 71 M'V, Figure 2-6, Table 2-2) is much less than that 
of human CA II (maximal k^/K^ 3 x 10 3 M'V 1 ; (Steiner et al., 1975)). 
Proton Transfer in CA VII 

Histidine 64. Several results suggest that His 64 is the predominant proton shuttle in 
CA VII, as it is in CA II. First, there is no other residue in the active-site cavity which is a 



43 
likely shuttle of pK a near 7. Second, inhibition by CuS0 4 of 18 exchange catalyzed by rMCA 
VII shows properties very similar to those observed for inhibition by CuS0 4 of HCA II in 
which His 64 is a proton shuttle (Tu et al., 1981). In HCA II cupric ion coordinates to the 
imidazole side chain of His 64 and blocks the proton transfer role of this residue (Eriksson 
et al., 1986). This results in inhibition of R H2C /[E] which is dependent on proton transfer, but 
has no effect on R[/[E] which measures interconversion of C0 2 and HC0 3 in the first stage 
of catalysis (equation 2-1). The same pattern is seen for rMCA VII. Third, the pH profile for 
rMCA VII in Figure 2-4 is nearly identical with that of human CA II. And finally, as in the 
case of CA II (Tu et al., 1989a), replacement of His 64 by Ala has removed a predominant 
proton shuttle in CA VII, this result is described below. 

The magnitude of Ka/^m f° r hydration was found to be unchanged between rMCA 
VII and the H64A mutant of rMCA VII (Figure 2-2; Table 2-2). Thus, the rMCA VII H64A 
mutant is not affecting the catalysis of the interconversion of CO, and bicarbonate at the zinc 
(equation 1-1, Chapter 1, page 9). Yet, there is an absence of one of the two ionizations in 
k^/K,,, near 6 that was observed in wild-type rMCA VII. This maybe interpreted as the 
removal of an influence on the ionization state of the zinc-bound water by His 64 upon 
mutation of the histidine at position 64 to alanine (see discussion beginning on page 46). In 
contrast, both k^ and Rh 20 /[E] are significantly reduced for H64A compared with wild-type 
(Figure 2-3 and 2-4; Table 2-2). Also, there is a loss of one of the two ionizations in k^, and 
the difference between the pH profile for R H20 /[E] for rMCA VII and for rMCA VII H64A 
(inset in Figure 2-4) shows the bell-shaped pH dependence consistent with the loss of a single 
proton shuttle of pK a 7.1 ± 0.1, and corresponding to a loss in proton transfer capacity 



44 
(Rroo/tE]) of (3.3 ± 0.3) x 10 5 s' 1 upon mutation of His 64 to Ala. By this argument, the 
maximum near pH 6 in the pH profile for the rMCA VII is due to the function of His 64 as 
a proton shuttle (Figure 2-4). These results suggest that the H64A mutant is affecting proton 
transfer and that histidine is the predominant proton shuttle at position 64 in CA VII. 

Effect of the amino-terminal residues . The maximal values of k^/K,,, and apparent 
pK/s are the same for rMCA VII and for the truncation mutant MCA Vllb (Table 2-2). Thus, 
the amino-terminus has no role in and is not affecting the steps of equation 1-1 (Chapter 1, 
page 9), the catalysis by CA VII of the conversion of C0 2 into bicarbonate at the zinc. 
However, two results suggest that the truncation is affecting proton transfer, the twofold 
decrease observed at pH 9.1 in k^, for hydration compared with full length (Table 2-2), and 
the decrease by an order of magnitude in R H2 o/[ E ] near P H 6 (Figure 2-5). Aronsson et al. 
(1995) observed catalysis by a truncated variant of human CA II in which 20 residues at the 
amino terminus were removed; it had an overall C0 2 hydration activity 1 5% of wild-type 
human CA II. 

Thus, it is reasonable to ask whether the truncation of the amino-terminal 23 residues 
has decreased the ability of His 64 to function as a proton shuttle. This suggestion is 
supported by the difference between the pH profile for R H2 c/[E] f° r rMCA VII and for the 
truncated form MCA Vllb shown in the inset in Figure 2-5. This difference plot shows the 
bell-shaped pH dependence consistent with the loss of a single proton shuttle of pK a 6.9 ± 0.2, 
and corresponding to a loss in proton transfer capacity (R H2 c/[E]) of 3.4 x 10 5 s" 1 upon 
truncation, similar to the removal of H64A from wild-type rMCA VII (see the difference plot 
in Figure 2-4). By this argument, the small maximum near pH 6 in the pH profile for the 



45 
truncated mutant MCA Vllb is due to the reduced capacity of His 64 to act as a proton 
shuttle (Figure 2-5). 

The full removal of His 64 in the truncation mutant (H64A MCA Vllb) results in a 
change in the pH dependence when compared to the truncation mutant alone (Figure 2-5). 
The difference in values of IWP] for MCA Vllb and H64A MCA Vllb is also given in the 
inset in Figure 2-5. Here again, the data are consistent with the loss of a single proton donor 
of pK a 7.5 ± 0.2 corresponding now to a smaller loss in proton transfer capacity of 5 x 10 4 



s- 1 . 



By this explanation the proton shuttle capacity of His 64 is lessened by removal of the 
amino-terminal 23 residues of rMCA VII. There are several possibilities for this loss. The first 
is a conformational change of the truncated form MCA Vllb in which His 64 is at a distance 
less effective for proton transfer or in which His 64 has a broader range of side-chain 
conformations than in the full-length enzyme and hence spends less time in the conformations 
appropriate for proton transfer. It is also possible that the three histidines of the amino- 
terminal 23 residues (Figure 2-1) are the significant proton donors in rMCA VII. This is 
unlikely because based on the structure of CA II (Eriksson et al., 1988b) the distance of these 
residues from the zinc (> 18 A) is much greater than for His 64 (~7 A) and therefore this 
distance would not be considered optimal for significant proton transfer; moreover, it is His 
64 that has been shown to play the predominant role as proton shuttle in rMCA VII as 
discussed above and in isozyme II. It is possible however, that these three histidines could 
play a role as secondary proton shuttles in a proton relay mechanism with His 64 since there 
is an obvious difference upon truncation of the H64A mutant (MCA Vllb H64A) when 



46 
compared to the full-length H64A (rMCA VII H64A) in the pH profiles of R H20 /[E] at pH 
near 6 (Figure 2-5). Interestingly, it is found that deletions of amino-terminal residues of 
carbonic anhydrases that lack an effective proton shuttle at position 64, namely CA III 2 and 
CA V (Heck et al., 1994), show catalytic properties nearly identical with their full-length 
counterparts. This finding may support the suggestion that the amino-terminal end of rMCA 
VII influences the function of His 64 as proton shuttle. This topic is under further study. 

Basic residues . Several results suggest there is another residue(s) in the active site of 
CA VII, in addition to His 64, that is participating as a proton shuttle in catalysis. In pH 
profiles of k^, for both wild-type and the H64A mutant there is an ionization at high pH 
(Table 2-2, Figure 2-2). Also, in pH profiles of R H2 o/[E] for full-length H64A and the 
truncated H64A mutant, where the capacity of His 64 to function as a proton shuttle group 
is removed, values of R H2C /[E] plateau near 3-6 x 10 4 s' 1 at pH > 8. This plateau at high pH 
represents proton transfer from a donor group of pK a > 9 to the zinc-bound hydroxide in the 
dehydration direction of catalysis (Figure 2-4 and 2-5). The identity of the proton shuttle 
residue(s) that is contributing to catalysis at high pH is one focus of my work in isozyme V 
and the results of that work will be discussed in Chapter 3. 
Assignment of pK 3 Values 

The pH dependence of k^/K,,, in the CO, hydration direction can be described by two 
ionizations for rMCA VII, near pK a 6.2 and 7.5 (Figure 2-2, Table 2-2). The pH rate profile 
for the esterase activity of rMCA VII also appears dependent on two ionizations with similar 
values of pK a (Figure 2-6, Table 2-2). One of these ionizations is clearly that of the zinc- 



Hevia, A., Tu. C. K., Silverman, D. N., and Laipis, P. J., unpublished observations 



47 
bound water; the other ionization most likely results from a perturbation of the pK a of the 
zinc-bound water caused by the electrostatic interaction of a nearby group. Perturbations on 
the pK a of the zinc-bound water have been described for mutants of CA. For example, in 
human CA II the introduction of a glutamate (Forsman et al., 1988) or a histidine (Behravan 
et al., 1991) in the active site changes the pH dependence of k^/K^ for ester hydrolysis from 
a single ionization to one described by two ionizations. 

By comparing data from steady-state measurements for wild-type and mutants in CA 
VII an assignment of these two ionizations in pH profiles of k^/K^ for rMCA VII has been 
achieved. Upon mutation of His 64 to Ala in rMCA VII the ionization near 6 disappears in 
pH profiles of k^/K,,,, leaving the second ionization near 7.5 (Figure 2-2; Table 2-2). This 
is also the case for the two truncation mutants of CA VII, MCA Vllb and MCA Vllb H64A 
where only an ionization near 7.5 persisted in pH profiles of k ca /K m (Table 2-2). These results 
suggest His 64 perturbs the pK a of the zinc to yield the second ionization near pH 6. 

The pH profile for k^ for hydration of CO, catalyzed by rMCA VII also contains two 
ionizations, one with a pK a of 6.2 ± 0.2 and one with a poorly determined pK a near 8.2 
(Figure 2-3; Table 2-2). One of these ionizations is His 64 and the other ionization should 
represent some other proton shuttle(s) in the active site. The pH profile for k cal for the H64A 
mutant is lacking an ionization at pH 6, and the ionization at high pH remains, which is in 
agreement with our assignment of the pK a value for His 64 near 6 in pH profiles of k^/K,,, as 
described above. The observations in rMCA VII of two ionizations influencing k^, for 
hydration are very similar to the observations of Hurt et al. (1997) in which k ca , for murine CA 
IV was found to depend on two ionizations. One of these had an apparent pK a of 6.3 and was 



48 
assigned to His 64 upon analysis of the H64A mutant, and a second with pK a of 9.1 may 
represent proton transfer from basic residues more distant from the zinc than His 64; this 
result is observed for other isozymes of CA and is extensively discussed in Chapter 3 
(Earnhardt et al., 1998b; Silverman et al., 1998). 

The bell-shaped pH profiles typically found for R H2 c/[ E ] can be described by equation 
2-4 which expresses Rh 2 c/[ e ] as th e product of the protonated form of the donor group and 
the unprotonated form of the aqueous ligand of the zinc (Silverman et al., 1993). By this 
analysis, the data of Figure 2-4 for unmodified rMCA VII demonstrate two values of pK a , 
near 5.8 and 7.4, which again suggest the same ionizations as observed in k^JK^ for C0 2 
hydration and hydrolysis of 4-nitrophenyl acetate (Table 2-2). However, the 18 0-exchange 
data of Figure 2-4 taken alone are equally consistent with the following two assignments: 1) 
The pK a of 7.4 is the zinc-bound water and the pK a of 5.8 is His 64 with k B = (1.3 ± 0.4) x 
10 7 s" 1 for the rate constant for intramolecular proton transfer in the dehydration direction (as 
shown for 18 exchange in equation 2-2); 2) The pK a of 7.4 is for His 64 and pK a of 5.8 is the 
zinc-bound water in which case k B = (3.2 ± 0.4) x 10 5 s" 1 . Assigning the lower ionization near 
6 to His 64 in the pH profile of R H 2c/[E] f° r rMCA VII yields a rate constant for proton 
transfer an order of magnitude higher than the two other fast isozymes, CA II and IV (Hurt 
et al., 1997; Tu et al., 1989a). Therefore, it is helpful to compare the 18 0-exchange data with 
the steady-state turnover for dehydration. Like k B , the turnover number for dehydration, k cat) 
is dependent on the proton transfer to the zinc-bound hydroxide. The maximal value of k cal 
for dehydration catalyzed by rMCA VII was 2 x 10 5 s" 1 (Table 2-2), a value consistent with 
assigning the pK a near 7.4 to His 64 and the pK a near 5.8 to the zinc-bound water. This 



49 
assignment of His 64 in pH profiles of R H2C /[E] to 7.4 is supported upon comparison to pH 
profiles of rMCA VII H64A and the truncation mutant MCA Vllb, both mutants are lacking 
a predominant proton shuttle and are missing the pK a of 7.4 and yield a pK a for the zinc- 
bound water nearer to 6. However, there is a discrepancy in assigning the pK a values of His 
64 and the zinc-bound water in RWtEL m that for steady state conditions it is the lower pIC, 
near 6.2 that represents His 64 as described in the previous paragraphs. Further experiments 
are needed to understand this difference and provide a better interpretation. 
Inhibition 

The inhibition of rMCA VII by the sulfonamides acetazolamide and ethoxzolamide 
measured by 18 exchange is greater than for the other isozymes (Table 2-1). These 
sulfonamide inhibitors are expected to bind directly to the zinc and adhere to the hydrophobic 
side of the active-site cavity as demonstrated in human CA II (Eriksson et al., 1988b). Many 
of the residues implicated in sulfonamide binding with CA II are conserved in rMCA VII. 
Two exceptions are L204 and C206 in CA II which in murine CA VII are serines. 
Conclusions 

Murine CA VII is a highly conserved isozyme of carbonic anhydrase and this 
conservation may be indicative of its functional importance. It is a very efficient carbonic 
anhydrase with catalytic activity 100-fold greater in k cat over the slowest CA isozyme III. It 
is therefore similar to the most active of the mammalian carbonic anhydrases, isozymes II and 
IV. Moreover, among these isozymes it is the most inhibited by two widely-used 
sulfonamides when measured by 18 exchange. This indicates a highly specific interaction of 
inhibitors with CA VII that should be pursued by X-ray crystallography of the ET complex. 



50 
Unique upon comparison to the other wild-type mammalian isozymes, CA VII demonstrates 
the effect of two ionizations in the pH profile of k^/K.. This suggests a close interaction 
between the zinc and His 64 that is missing in other carbonic anhydrases and may contribute 
to rapid proton transfer. One of these is the ionization of the zinc-bound water (pK, 7.5) and 
the second is suggested to be His 64 (pK a 6.2). Moreover, for the first time a role for the 
amino-terminal end in enhancing proton transfer has been determined in catalysis for a 
carbonic anhydrase. More specifically, the amino terminus may be restricting His 64 to useful 
conformations for proton transfer. 












CHAPTER 3 

INTRAMOLECULAR PROTON TRANSFER FROM MULTIPLE SITES IN 

CATALYSIS BY MURINE CARBONIC ANHYDRASE V 



Introduction 

Carbonic anhydrase V (CA V) is a mitochondrial enzyme found predominantly in liver 
(Table 1-1, Chapter 1, page 2); it is a member of the a class of carbonic anhydrases which 
includes the mammalian isozymes (reviewed in Dodgson, 1991). The catalytic properties of 
murine carbonic anhydrase V (MCA V) 3 have been characterized (Heck et al., 1994) and its 
crystal structure is determined to 2.45-A resolution (Boriack-Sjodin et al., 1995). Similar to 
the other carbonic anhydrases of the a class, MCA V is a monomelic zinc metalloenzyme of 
molecular mass near 30 kDa that catalyzes the hydration of carbon dioxide to form 
bicarbonate and a proton. The catalytic pathway of MCA V is similar to that of the well- 
studied C A II in many respects. The two stage catalysis of these isozymes is described in 
detail in Chapter 1 (page 9). 

For CA II, proton transfer proceeds through an intramolecular proton shuttle 
(designated in equation 1-2 as H + to the left of E, Chapter 1, page 9) which subsequently 
releases the proton to solution. In CA II this intramolecular proton shuttle has been identified 



Abbreviations: MCA V, murine CA V; HCA II, human carbonic anhydrase II; Y64A, the 
mutant with Tyr 64 replaced by Ala; Tris, tris(hydroxymethyl) amino methane; Ted, 1,4 
diazabicyclo [2.2.2] octane or triethylenediamine. 

51 



52 
as His 64 (Steiner et al., 1975; Tu et al., 1989a) which extends into the active site cavity with 
the N5 of its imidazole ring 8.2 A from the zinc and with no apparent interactions with other 
residues (Eriksson et al., 1988b). MCA V has a tyrosine residue at position 64 which is not 
an efficient proton shuttle (Heck et al., 1994). However, it is possible to activate MCA V by 
placing a histidine residue at position 64 along with other changes in the active site (Heck et 
al., 1996). Studies using isotope effects, pH dependencies, and chemical rescue have shown 
that these intramolecular proton transfer steps are rate-determining for maximal velocity 
(Silverman and Lindskog, 1988; Tu et al., 1989a). 

Position 64 in carbonic anhydrase is not the only site from which proton transfer can 
occur. Liang et al. (1993b) placed a histidine residue at four other positions within the active- 
site cavity of isozyme II and found that a His at 67 was capable of enhancing catalytic activity 
but at a level no greater than 20% of the wild-type enzyme. Ren et al. (1995) showed that His 
67 in human CA ITI is capable of proton transfer but again not as efficiently as His 64 in CA 
III. It is significant that the mutants of carbonic anhydrase lacking a histidine proton shuttle 
in the active-site cavity can still sustain a catalytic turnover k cal for hydration at pH near 9 of 
10 4 s l as for human CA III (Jewell et al., 1991) and as great as 3 x 10 5 s" 1 for MCA V (Heck 
et al., 1994). This suggests the presence of one or more basic residues that act as proton 
shuttles. 

Since MCA V supports catalysis at a rapid rate at high pH, but is lacking His 64 as 
a prominent proton shuttle, it is pertinent to ask what residues in MCA V support this 
significant activity. There are a number of lysine and tyrosine residues in MCA V located in 
the active-site cavity or around its rim. In this study these residues have been replaced with 



53 
alanine and the initial velocities in catalysis by the resulting mutants were measured using 
stopped-flow spectrophotometry; catalysis of 18 exchange between C0 2 and water was also 
measured using mass spectrometry. The results show that the catalytic activity in MCA V is 
supported by multiple proton transfers involving a number of ionizable groups of basic pK a , 
some more distant from the zinc than residue 64. Although there is no single prominent 
proton shuttle, Lys 91 and Tyr 13 1 with their amino and phenolic hydroxyl groups 14.4 A and 
9.1 A from the zinc, as shown in Figure 3-1 (Boriack-Sjodin et al., 1995), account for about 
half of the catalytic turnover. Moreover, the interaction between these proton shuttles in 
catalysis is not simply additive, but antagonistic reflecting their adjacent location and 
suggesting a cooperative behavior in facilitating the proton transfer step of catalysis. 
Replacing four of these possible proton shuttle residues produced a multiple mutant that has 
10% of the catalytic turnover k cal of the wild-type, suggesting that the main proton shuttles 
have been accounted for in MCA V. As a control, the replacements were determined to cause 
relatively small changes in k a J¥i m for hydration which measures the interconversion of C0 2 
and HC0 3 ~ in a stage of catalysis that is separate and distinct from the proton transfers. 

Materials and Methods 

Site-Specific Mutagenesis 

The coding sequence of CA V was derived from BALB/C mouse liver mRNA by 
reverse transcription and PCR (Heck et al., 1994; Heck et al., 1996). The mutant forms of 
MCA V used in this study were prepared by Dr. Minzhang Qian in the laboratory of Dr. Philip 
J. Laipis, and were created using a mutating oligonucleotide (Kunkel, 1985) in the pET31 



54 




Figure 3-1. The location of ionizable residues near the active site cavity of murine carbonic 
anhydrase V from the crystal structure of Boriack-Sjodin et al. (1995). The three ligands of 
the zinc are His 94, 96, and 1 19. 



55 
expression vector system (Tanhauser et al., 1992); alterations were verified by DNA 
sequencing. 
Expression and Purification 

Wild-type and mutant forms of the enzyme were expressed from the pET vector after 
transformation into E. coli BL21(DE3)pLysS (Studier et al., 1990). All of the expressed 
enzymes were truncated forms lacking the first 5 1 amino-terminal residues. In a sequence 
numbering scheme consistent with CA II, the expressed MCA V variants began at residue 22, 
Ser. This truncated form of MCA V (denoted MCA Vc by Heck et al. (1994)) has been 
shown to have identical catalytic properties to MCA V expressed from both a full length 
coding sequence and a 30 residue truncation of MCA V (Heck et al., 1994). 

Purification was performed through previously described procedures with slight 
modifications (Heck et al., 1994). Frozen cells containing expressed recombinant MCA V 
mutants were thawed in a solution of 25 mM Tris pH 8.5 containing 2 mM EDTA 0.2 mM 
phenylmethylsulfonyl fluoride, 5 mM benzamidine, 0.4 mM MgCl 2 , 0.4 mM ZnS0 4 , 0.1% P- 
mercaptoethanol, 0.1 mg/mL deoxyribonuclease I and 0.5 mg/mL lysozyme with stirring for 
two hours in 4°C. After cell lysis the cell debris was pelleted by centrifugation at 43000 x g 
for 15 minutes at 4°C. The supernatant was added to a gel filtration column (Ultrogel AcA 
44, LKB). The protein eluate was then subjected to affinity chromatography on a gel 
containing p-aminomethylbenzenesulfonamide coupled to agarose beads as described by 
Khalifah et al. (1977). Enzyme purity was determined as in Chapter 2 (page 26). Active MCA 
V mutant enzyme concentration was determined by inhibitor titration of the active site with 
ethoxzolamide by measuring 18 exchange between C0 2 and water (see below). The enzyme 
was then stored at 4°C. 



56 
Stopped-flow Spectrophotometry 

Initial velocities were determined by following the change in absorbance of a pH 
indicator (Khalifah, 1971) at 25 °C using a stopped-flow spectrophotometer (Applied 
Photophysics Model SF. 1 7MV). This method is described in detail in Chapter 2 (page 26) 
and was used with only minor variations. Here, the C0 2 concentrations for the substrates 
ranged from 0.7 to 17 mM. The buffer-indicator pairs, pK a values and the wavelengths 
observed are also described in Chapter 2 (page 26) with two additions as follows: 1,2 
dimethyl imidazole (pK, = 8.2) and m-cresol purple (pK_, = 8.3) 578 nm and Ted (pK, = 9.2) 
and thymol blue (pK a = 8.9) 596 nm. 
lg O Exchang e 

An Extrel EXM-200 mass spectrometer utilizing a membrane permeable to gases was 
used to measure the rate of exchange of l8 between species of C0 2 and water catalyzed by 
the carbonic anhydrases (equations 2-1 and 2-2, Chapter 2, page 27; Silverman, 1982). 
Experiments were at 25 °C. No buffers were added except where indicated, and a total ionic 
strength of 0.2 M was maintained with Na 2 S0 4 . Solutions contained 25 mM total substrate 
([COJ + [HC0 3 ']). This method is described in detail in Chapter 2 on pages 27 through 29. 
Values of k^/K^ obtained from l8 0-exchange techniques for the mutants of Table 3-1 were 
determined by Dr. C. K. Tu (University of Florida). Solvent hydrogen isotope effect 
experiments are as described on page 29 in Chapter 2. 

Results 

The mutants constructed for this study have the potential proton transfer residues at 
positions 64, 91, 131, 132 and 170 replaced with alanine in single and multiple mutations 



Q. 






o 

Ei 

e_ 

X 

a. 

— 

'5 

E 
o 



<u 

c 

"3 

■*-» 
x> 

o 

u 

> 

u 

x: 



Ji 

T3 

C 
<S 

C 

i- 

T3 

>^ 

X 

o 






^ 



M 

^ " c 
O O 

« -g 

> g 
J j 



■ • c 

•2 "° 



S 
a. 



t/5 



a. 



a. 



3. 



J3 



u 



m 






— 


(N 




CS 


O 


(*• 


m 


o 


o 


*/"! 


o 


-H 


00 


00 


-H 


-H 


00 


-H 


r» 


A 


A 


NO 


on 


A 


M 


00 






00 


00 




00 



-H 

o 

00 



rs 



o 
-H 

00 



o 


O 


o 


o 


^— 


i— 


1 — 1 


-H 


-H 


-H 


■H 


o 


o 


o 


o 



o 
■H 



-H -H 
On (N 
— (N 



O 
-H 
NO 



-H 



-H -H 

o o 
p-l f- 







M 


(N 


— 


— 


(N 


o 


O 


o 


O 


O 


o 


O 


■H 


-H 


■H 


■H 


■H 


-H 


■M 


rN 


00 


— 


tN 


o 


(S 


<-~ 


ON 


00 


On 


On 


ON 


On 


On 



O 


O 


o 


o 


O 


O 


-H 


-H 


-H 


-H 


-H 


-H 


O 


O 

00 


o 


O 


O 


o 

00 



m 



O 

-H 



o o 
-H -H 
in — ■ 



-H -H 









-H 



o 
■H 

On 



NO 







< 














^^ 














r^ 














*■— 












< 


£ 
^ 








< 
— 




CI 


i— 








m 




^^ 


On 


<u 










> 


\Z t 


a. 






s 


< 


^ 


<> 




< 


< 


5 


Tt 


•<t 


t 


T3 


^-* 


fl 




NO 


NO 


NO 


772 


On 


. — i 


On 


> 


> 


> 


£ 


M 


> 


* 



o 

-H 

NO 



-H -H 

NO 00 



o 
-H 

00 



o 

CM 

-H 
o 



o 

-H 



-H 



< 

(N 



in © 

oo -h 

A >n 

ob 



O 

r*i o 

-H -H 



O 



o 
-H 



-H 

NO 



O 



X U 



o 

NO 



o 


O 


o 


-H 


-H 


-H 


— 


On 


^N 


On 


00 


On 



O 


O 


o 


NO 


-H 


-H 


-H 


-H 


O 


O 
r- 


o 


<^1 



NO 



ON 



3" no 

On On 
On On 



<S aj 



1) SJ 'Si 



58 



10' 










t 

10 4L 



7 
pH 



8 



Figure 3-2. The pH dependence of k^/Km for hydration of C0 2 determined by 18 exchange 
catalyzed by (•) wild-type MCA V (Heck et al., 1994); (O) K91 A/Y13 1 A MCA V; and (A) 
Y64A/K91 A/Y131A MCA V at 25 °C. Total ionic strength was maintained at 0.2 M by 
addition of Na 2 S0 4 ; no buffers were used. Lines are a nonlinear least-squares fit to a single 
ionization for wild-type and Y64A/K91A/Y131A and to two ionizations for the 
K91A/Y131A mutant resulting in the parameters given in Table 3-1. 



59 
(Figure 3-1, Table 3-1). The pH profiles for k c- /K B for hydration of C0 2 , determined from 
18 0-exchange rates, varied as if dependent on the base form of a single ionizable group 
(Figure 3-2) with apparent values of pK, that were similar for each mutant, from pK, 7. 1 to 
7.9 (Table 3-1). The double mutant K91A/Y131A MCA V was an exception in that two 
ionizations fit the data better than one; however, the apparent values of pK a were very similar 
for the two ionizations (Table 3-1). The alanine replacements did not cause large changes in 
the maximal values of k^/K,,, which had magnitudes ranging from 1.9 to 6.0 x 10 7 M'V for 
all of the mutants studied (Table 3-1). The magnitudes of KiJ^m appeared to occur in three 
groups as shown in Table 3-1; for example, the mutants containing Tyr 64 including wild-type 
MCA V have values of k^/K^ very near 3.5 x 10 7 M'V and the mutants containing Ala 64 
have values very near 2.0 x 10 7 M'V 1 (Table 3-1). The ratio k^/Kg, contains rate constants 
for the conversion of C0 2 into HC0 3 ' up to and including the first irreversible step, the 
departure of HC0 3 ' (equation 1-1, Chapter 1, page 9). Therefore, these mutants are not 
causing significant changes in this stage of catalysis or influencing the surrounding 
environment of the zinc-bound water in a manner that would alter its pK a or catalytic activity. 

Measurement by stopped-flow of the steady-state constants k^, for C0 2 hydration had 
a pH dependence that could also be fit to a single ionization with apparent values of pK a in 
the narrow range of 8.8 to 9.2 for wild-type and mutants (Table 3-1) with typical data shown 
in Figure 3-3. The variation in the maximal values of k cat ranged from 3.2 x 10 5 s" 1 for the 
wild-type enzyme to 0.32 x 10 5 s" 1 for the multiple mutant Y64A/K91A/Y131A/K132A, 
which is 10% of k^, for the wild-type (Table 3-1). 

The 18 0-exchange rate constant R H2 c/[E] describes the rate of release of H 2 18 from 
the enzyme into solvent water at chemical equilibrium and involves proton transfer to the 






60 



ft 




Figure 3-3. The pH dependence of k,.,, for hydration of C0 2 determined by stopped-flow 
spectrophotometry catalyzed by (•) wild-type (Heck et al., (1994)); (O) K91A/Y131 A MCA 
V; and (A) Y64A/K91A/Y131AMCA V at 25 °C. Total ionic strength was maintained at 0.2 
M by addition of NajSCv Lines are a nonlinear least-squares fit to a single ionization resulting 
in the parameters given in Table 3-1 . 



61 
zinc-bound hydroxide in the dehydration direction (equation 2-1, Chapter 2, page 27). The 
pH profiles of R^c/tE] were typically bell-shaped for the mutants of Table 3-1 and yielded 
a rate constant k B for intramolecular proton transfer and values of pK a for the zinc-bound 
water and for the intramolecular proton donor determined by least-squares fitting of equation 
2-4 (Chapter 2, page 29) to the pH profiles of R mc J[E]. The values of the pK, of the zinc- 
bound water were in agreement, within 0.1 or 0.2 units, with those obtained from the pH 
profiles of lwK,. Application of these procedures to typical data are shown in Figure 3-4 for 
wild-type MCA V and K91 A/Y13 1 A MCA V. The 18 0-exchange results again show a narrow 
range of values for the proton donor of pK, from 8.3 to 8.9 (Table 3-1). These values confirm 
the pK, near 9 found from the pH profile of k^, for the proton shuttle also shown in Table 3- 
1 . The wild-type enzyme had the largest value of the rate constant k B for intramolecular 
proton transfer with the smallest value being observed for the triple mutant 
Y64A/K91A/Y131A at 17% of the wild-type (Table 3-1). The magnitudes of k B and k cal can 
only be compared qualitatively since they represent proton-transfer in the dehydration and 
hydration directions. 

The solvent hydrogen isotope effect (SHIE) observed for catalysis of CO, hydration 
by Y64A/K91A/Y131A MCA V was 1.4 ± 0.2 for k^/K^ at pH 9.2. This is consistent with 
no rate-contributing proton transfer in the interconversion of CO, and HC0 3 " (equation 1-1, 
Chapter 1, page 9). The SHIE at pH 9.2 on k^,, was 4.1 ± 0.5, consistent with rate- 
determining proton transfer involving the aqueous ligand of the zinc (equation 1-2, Chapter 
1, page 9). 

The capacity of mutants of MCA V to be enhanced in catalysis by proton donors 
from solution through chemical rescue was also measured by 18 exchange for the 



62 



10 



Gfl 



W 10 4 



o 

DC 

PS 




Figure 3-4. The pH dependence of Rfco/PL tne rate constant for release of 18 0-labeled 
water from the enzyme, catalyzed by (•) wild-type MCA V and (O) the mutant 
K91A/Y131 A MCA V at 25 °C. The total concentration of all species of C0 2 was 25 mM, 
the total ionic strength of solution was maintained at 0.2 M by addition of NajSO^ and no 
buffers were added. 



63 
dehydration direction of catalysis. These experiments were done using the buffers of small 
size, imidazole and 1,2-dimethyl imidazole. It was found that imidazole was able to activate 
R H20 /[E] catalyzed by Y64A/F65A MCA V in a saturable manner at pH 6.3 (Figure 3-5). 
Imidazole had a slight inhibitory effect on R[/[E] and caused these values to decrease with an 
apparent K^ listed in the legend to Figure 3-5. The maximal value of R^o/fE] when corrected 
for this inhibition was identical to the value for Y64H/F65A MCA V in the absence of buffer 
at pH 6.3 (Table 3-2; Figure 3-5). Similar results were obtained through chemical rescue of 
the triple mutant, Y64A/K91 A/Y13 1A, with 1,2-dimethyl imidazole. R H2 o/[E] catalyzed by 
this triple mutant increased in a saturable manner upon addition of 1,2-dimethyl imidazole at 
pH 8.2 (Table 3-2). The maximal value of R H2C /[E] is very similar to that measured in the 
absence of buffers with Y64H/F65A MCA V at the same pH (Table 3-2). As a control, 1,2- 
dimethyl imidazole catalyzed by Y64A/K91A/Y131A caused no change in R,/[E] at 
concentrations up to 200 mM. 

The capacity of mutants of MCA V to be enhanced in catalysis by proton acceptors 
from solution through chemical rescue was also measured under steady-state conditions. The 
estimated values of k^ for C0 2 hydration catalyzed by Y64A/K91A/Y131A MCA V 
increased from a value close to 1 x 10 3 s' 1 to values approaching (1.0 ± 0.3) x 10 5 s' 1 as 
concentrations of 1 ,2-dimethyl imidazole increased from 1 to 200 mM (pH 8.2, 25 °C, ionic 
strength maintained at a minimum of 0.2 M; Data not shown). Again, this is close to the value 
of k cat (2.0 ± 0.4) x 10 5 s" 1 under these conditions for Y64H/F65A MCA V (Heck et al., 
1996). 

Chemical rescue experiments measured at chemical equilibrium by the 18 0-exchange 
technique requires buffers of small size to fit in the active site and directly shuttle protons to 



64 




50 100 150 

[Imidazole] mM 



200 









Figure 3-5. The dependence of R^c/IE] (•) and R^FE] (□) as a function of the 
concentration of imidazole at pH 6.3 and 25 °C. The total concentration of all species of C0 2 
was 25 mM and the total ionic strength of solution was maintained at a minimum of 0.2 M 
by addition of NajSO^ The data for R^c/Pi] approach a maximal value of (1.1 ±0.2)x 10 5 
s" 1 when corrected for the apparent inhibition manifested in R,/[E] (Kj of 0.40 ±0.17 M). The 
apparent K^ for this buffer was near 91 mM. 






65 



Table 3-2: Catalysis Enhanced by Proton Donors from Solution in Mutants of MCA V: 
Values of R H20 /[E], the Rate of Release of I8 0-labeled Water from the Active Site. 

Mutant pH R H 2c/[E] 

(xlQ-V) 

Y64H/F65A a 6.3 1 

8.2 0.2 

Y64A/F65A with 

imidazole" 6.3 1.1 ±0.2 

Y64A/K91A/Y131A 

with 1,2 dimethyl imidazole 1 * 8.2 0.2 ±0.3 

Note : All data obtained at 25 °C with a minimum ionic strength maintained at 0.2 M with the 

addition of Na 2 S0 4 . 

a C. K. Tu and D. N. Silverman, unpublished data. 

b Maximal values of R^o/fE] were determined from the dependence of Rh 20 /[E] as a function 

of the concentration of buffer in catalysis by MCA V Y64A/F65A at pH values equal to pK a 

and under the experimental conditions described in the legend of Figure 3-5. The data for 

R-H2c/[E] were corrected for the apparent inhibition manifested in R,/[E]. 



66 
the zinc-bound hydroxide (Tu et al., 1990). Under these conditions, the bulky sized buffer 
Ted caused no change in R^P] catalyzed by the triple mutant Y64A/K91 A/Y13 1 A (at pH 
9.3 and 25 °C) up to concentrations of 100 mM. As a control, Ted was also found to cause 
no change in kj}^ catalyzed by Y64A/K91 A/Y13 1 A MCA V up to similar concentrations. 
This result indicates that Ted cannot directly donate protons to the zinc-bound hydroxide and 
therefore, does not fit in the active site. However, under steady state conditions where buffer 
in solution is the final proton acceptor in each cycle of catalysis, chemical rescue of the triple 
mutant Y64A/K91 A/Y13 1 A MCA V was achieved by addition of similar concentrations of 
the buffer Ted as used in the 18 0-exchange experiments. In these experiments, Ted produced 
an increase of greater than 10-fold in the initial velocity of CO, hydration. As will be 
discussed, the comparison of the results of chemical rescue experiments from 18 0-exchange 
and steady state techniques provides evidence that the proton transfer groups are located near 
the surface of MCA V and not buried in the active site. 

Discussion 

The aim of this study is to identify in MCA V the residue or residues of pK a near 9 
that act as proton acceptors in the hydration direction of catalysis. For this purpose the 
proton transfer capacity was investigated for a number of lysine and tyrosine residues in the 
active-site cavity and near its rim. Five potential proton shuttle residues in the active-site 
cavity or near its mouth in MCA V were replaced; this includes many evident basic groups 
in the vicinity of the active site that could participate as proton shuttles. These have been 
replaced by alanine and the effect on k cat for hydration and on 18 exchange have been 
measured. 



o / 

Among the variants of MCA V in Table 3-1, wild-type MCA V has the largest value 
of k^, and the various single mutants with potential proton shuttles replaced have lower 
values. A similar observation is also made for k B , a rate constant for intramolecular proton 
transfer determined from 18 exchange, in which again the wild-type enzyme has the largest 
value (Table 3-1). These decreases in k ca , for hydration and k B observed in our mutants of 
MCA V are interpreted as decreased efficiency of the intramolecular proton transfer 
processes. This interpretation is supported by the following considerations. A wide body of 
previous data indicates that k^, for catalysis by carbonic anhydrase is dominated by 
intramolecular proton transfer between the zinc-bound water and proton shuttle residues 
(Steiner et al., 1975; Tu et al., 1989a; Lindskog, 1997), and the studies with murine CA V 
are also consistent with such rate-determining steps for k cal (Heck et al., 1994; Heck et al., 
1996). Many of the previous experiments to confirm rate-limiting proton transfer in this 
catalysis have been repeated in this study with one of the least active mutants and other 
mutants of Table 3-1 : Y64A/K91 A/Y13 1 A has a solvent hydrogen isotope effect of 4. 1 on 
k eal for hydration; k„, is activated by proton donors in solution, an example of chemical 
rescue; and k^ and Rh 2C /[E] have pH profiles that are consistent with intramolecular proton 
transfer. 

There is no single replacement in Table 3-1 that causes a decrease in k cat or k B as large 
as the 40-fold decrease that resulted from the replacement of His 64 by Ala in HCA II (Tu 
et al., 1989a). Thus, it appears that in MCA V there is not one predominant proton shuttle 
group as in HCA II but many shuttle groups. The closest potential side chain among those 
studied is Tyr 64 with its hydroxyl oxygen 7.7 A from the zinc. However, this side chain is 



68 
pointing away from the zinc in the crystal structure and appears limited in its mobility by the 
adjacent Phe 65 which most likely accounts for the very slight reduction (or no change, Table 
3-1) when it is replaced by alanine (Heck et al., 1994; Heck et al., 1996). Among the 
remaining lysines and tyrosines of Table 3-1, the closest to the zinc are Tyr 131 with its 
hydroxyl oxygen 9 A from the zinc and Lys 91 with its NC 14 A from the zinc (Boriack- 
Sjodin et al., 1995). Accordingly, these replacements among the single mutants caused the 
largest decreases in k^, (Table 3-1). The residues Lys 132 and Lys 170 have their NC a 
distance of 19 A and 20 A respectively from the zinc; the decreases in k^, when these residues 
are replaced by Ala are very small (or no change) compared with k^, for wild-type (Table 3- 

1). 

The rate constants for intramolecular proton transfer k B determined by 18 exchange 

measure proton transfer in the dehydration direction and are different from the values of k ca , 

for hydration discussed above. The values of k B add an important component to 

interpretations of this work since they represent data taken in the absence of buffer, and unlike 

k^ measured at steady-state, do not contain the possibility of direct proton transfer between 

the buffer and the zinc-bound water. The decreases in k B for the mutants of Table 3-1 

compared with wild-type in general mirror the decreases in k cal . However there are notable 

exceptions such as for Y13 1 A which has a rather greater effect on k^, than on k B compared 

with wild-type (Table 3-1). On the other hand, K91 A has a greater effect on k B . No additional 

evidence is available to explain these observations. 

Hence, among the replacements of basic groups in Table 3-1 resulting in single 

mutants, the replacements of Lys 91 and Tyr 131 caused significant decreases in k cal for 



69 
hydration. This evidence is consistent with proton shuttle roles for Lys 91 and Tyr 131 as they 
participate in the intramolecular proton transfer steps. There are other explanations for these 
observations, but they can be considered less likely. For example, it is possible that these 
residues are not proton shuttles themselves but are residues that contribute to proton transfer 
by their effects on the formation of hydrogen-bonded water networks in the active-site cavity. 
The following observations indicate that changes in catalysis by the mutants of Table 3-1 are 
not significantly affected by any changes in such water structure. 

First, the lack of a significant effect of the replacement of the suggested proton 
transfer residues (K91, Y131) on kJK,, (Table 3-1) compared to wild-type suggests that the 
chemistry of C0 2 hydration at the zinc is not affected by replacements at these distant sites, 
possibly including changes in water structure. However, this approach needs further support 
since studies of human CA II have shown that the insertion of bulky residues including Phe 
at position 65 adjacent to the proton shuttle residue His 64 alters water structure in the active 
site of the crystal structure, an effect which is accompanied by significant decreases in k^, 
with smaller or no changes in kj¥^ (Scolnick and Christianson, 1996; Jackman et al., 1996). 
The crystal structures of MCA V and human CA II are very similar with backbone 
conformations that are superimposable with a rms deviation of 0.93 A (Boriack-Sjodin et al., 
1995). When the substitution Phe 65 to Ala is made in MCA V, a decrease in k^, is not 
observed for the resulting F65A mutant compared with wild-type (Table 1 in reference Heck 
et al., 1996). This suggests that the proton-transfer dependent values of k cal for MCA V, 
presumably involving proton transfer from more distant sites, are not affected by changes in 
water structure caused by the replacement Phe 65 to Ala. 



70 
Another observation suggesting that changes in water structure are not significantly 
involved in the data of Table 3-1 is that chemical rescue of certain of these mutants of MCA 
V with imidazole or 1,2-dimethyl imidazole activates catalysis to levels found for the mutant 
Y64H/F65A containing an unhindered imidazole as proton shuttle (Heck et al., 1996). Thus, 
the mutant Y64A/F65A when enhanced with imidazole achieved values of R H2C /[E] near 1 
x 10 s s' 1 (Table 3-2; Figure 3-5) identical to that of Y64H/F65A. Similarly, the mutant 
Y64A/K91 A/Y131A was activated by 1,2-dimethyl imidazole to levels of R H2C /[E] and k caI 
for hydration similar to that of Y64H/F65A in the absence of this buffer (or at very low buffer 
concentration). These observations also suggest that substitution of K91 and Y131 on the 
periphery of the active-site cavity have no measurable effect on proton transfer and 
presumably water structure when the proton shuttle is 1,2-dimethyl imidazole. 

Although Table 3-1 reports decreased catalysis upon replacement of Tyr 131 with Ala, 
there is the following evidence that proton transfer to enhance catalysis can occur from 
position 131. Chemical modification of Y131C MCA V with 4-chloromethyl imidazole and 
4-bromoethyl imidazole caused up to threefold enhancement of R H2C /[E] at pH < 7 with pH 
profiles consistent with the presence of a proton donor of pK a near 6 (Earnhardt et al., 
1998c). These results indicate that the imidazole group of the chemically modified Cys 131 
promotes proton transfer and shows that a proton shuttle at this site can act as a proton donor 
in catalysis. Attempts to observe an enhancement of catalysis with Y131H were not 
conclusive. 

Experiments were performed to eliminate some additional considerations as 
contributing to proton transfer, showing they have no significant effect on k cal . For example, 



71 
rate constants for the initial velocities of catalysis do not increase with an increase of enzyme 
concentration (Data not shown). Thus, there is no significant intermolecular proton transfer 
involving ionizable residues on the surface of other carbonic anhydrase molecules in solution. 
Such a possibility is unlikely due to the sub-micromolar concentrations of enzyme used in all 
of our experiments. 

The data of Table 3-1 indicate that K91 and Y131 make substantial contributions to 
proton transfer during hydration, but their replacement still leaves considerable activity, near 
3 x 10 4 s" 1 at pH near 9 for the quadruple mutant of Table 3-1, which indicates that there 
remain other proton shuttle residues. Although this is a high rate of catalysis, it is pertinent 
that proton transfer to hydroxide in solution could be close to this value; k 2 [OH"] ■ 
(10 9 M" 1 s' 1 )(10' 5 M) = 10 4 s' 1 at this pH, where k 2 is a roughly estimated diffusion-controlled 
bimolecular rate constant for hydroxide ion encounter with carbonic anhydrase perhaps 
similar to that found for cyanide (Prabhananda et al., 1987). Although hydroxide might 
contribute as a proton acceptor, the observation of a plateau in k cal at high pH (Figure 3-3) 
indicates that hydroxide is not the main proton acceptor at pH up to 9. The results indicate 
that the predominant proton shuttle residues, but not all of the proton shuttle residues, have 
been accounted for in MCA V. That the remaining catalytic activity in C0 2 hydration of our 
least active mutants of Table 3-1 still have a pK a of k^, near 9 indicates that the remaining 
proton shuttles are likely basic residues such as Tyr 58 or perhaps Lys 133 or even more 
distant basic groups. These basic groups are likely to have a thermodynamic advantage as 
proton acceptors compared with histidine residues of expected pK a near 6 or 7; besides, the 
crystal structure of the truncated form of MCA V used in these studies shows no histidine 



72 
residues near the mouth of the active-site cavity (Boriack-Sjodin et al., 1995). There is a 
further argument that the proton acceptors unaccounted for lie on the surface of the enzyme 
rather than deeper in the active-site cavity. That addition of the buffer Ted caused no 
enhancement of RWtE] catalyzed by Y64A/K91A/Y131A MCA V indicates that this bulky 
buffer cannot enter the active-site cavity to transfer a proton to the zinc-bound hydroxide ~ 
the catalysis is sustained by the various surrounding proton donors (and water) that are at 
their equilibrium protonation states in this isotope exchange at chemical equilibrium. 
However, Ted caused a very large increase in the initial velocity of catalyzed CO, hydration 
suggesting that it can accept protons from proton shuttle sites closer to or on the surface of 
the enzyme. 

Although in single mutants these sites had small changes compared with wild-type, 
the multiple mutants Y64A/K91 A/Y131A and Y64A/K91A/Y13 1 A/K132A had values of k cat 
reduced to 22% and 10% of that of wild-type while showing no substantial decrease in k^/K,,, 
(Table 3-1). This suggests that most of the significant proton shuttles of MCA V have been 
accounted for and emphasizes that there is no single prominent shuttle as in HCA II, but that 
a group of residues near the rim of the active-site cavity each make a relatively small 
contribution to the proton transfer to solution. 

The interaction between Lys 91 and Tyr 131 in the catalytic pathway is clearly not 
additive as indicated by comparison of k cal for these single mutants and the double mutant 
K91 A/Y13 1 A (Table 3-1). That is, these residues are not acting independently in their role 
supporting proton transfer. Rather the double mutant causes no additional decrease in 
catalysis beyond either of the single mutants. This is a form of antagonism, as described by 



73 
Mildvan et al. (1992), between two residues that becomes evident upon observing catalysis 
by the double mutant. Two possible explanations account for this antagonistic effect: 1) The 
side chains of Lys 91 and Tyr 131 are adjacent to one another at the mouth of the active site 
cavity (Figure 3-1). These two residues form a proton transfer chain in which both are 
required sequentially to transfer protons out to solution. 2) The antagonistic effect could be 
structural in which one residue is restricting the mobility of the second to conformations in 
which proton transfer occurs. 

As anticipated, this effect of basic residues is not specific for MCA V. The pH profiles 
of k^ for at least five of the seven functional isozymes in the a class, CA II, III, IV, V, and 
VII, demonstrate a dependence on ionizations at high pH which cannot be attributed to His 
64 (Silverman et al., 1998). For human CA II this is demonstrated in H64A (Silverman et al., 
1998). In human CA III there is an increment in k cal of unknown source observed at pH > 8 
(see Figure 2 of Jewell et al., (1991)). Murine CA IV wild-type has a pH dependence of k cat 
described by two ionizations, and in the H64A mutant the groups with pK a near 9 remain (see 
Figure 6 of Hurt et al., (1997)). Finally, as described in Chapter 2, the murine form of isozyme 
VII has a pH dependence for k cal described by two ionizations, one of which is the histidine 
at position 64 and the other at higher pH is proposed to be another active site residue(s) 
ionizing at high pH (Earnhardt et al., 1998a). It is possible to consider this common 
observation for many isozymes of carbonic anhydrase as due to an accumulation of basic 
amino acids occurring near the active site cavity in many of these isozymes; for example, CA 
II, III, IV, V and VII all contain a lysine at positions 169/170 as well as other lysines and 
tyrosines located at the mouth of the active site cavity. Thus, in conclusion, in MCA V and 






74 
likely in other isozymes of the a class of the carbonic anhydrases there are multiple proton 

transfers contributing to the overall catalytic efficiency of catalysis. 












CHAPTER 4 

CATALYSIS BY MURINE CARBONIC ANHYDRASE V IS ENHANCED 

BY EXTERNAL PROTON DONORS 



Introduction 

Chapter 3 demonstrates that the catalytic activity of MCA V is supported by a number 
of ionizable residues of basic pK a that act as proton shuttles. Such residues include Lys 91 
and Tyr 131 with their amino and phenolic hydroxyl groups 14.4 A and 9. 1 A from the zinc 
(Boriack-Sjodin et al., 1995) and they account for about half of the catalytic turnover 
(Chapter 3). This finding of multiple proton transfer residues is in sharp contrast to previous 
studies of CA II that emphasized a single proton shuttle, His 64, that sustains catalysis with 
a maximal turnover of 10 6 s" 1 (Silverman and Lindskog, 1988; Tu et al., 1989a). 

Placing histidine residues at strategic positions in the active site of carbonic anhydrase 
results in enhancement of catalytic rates for C0 2 hydration, some of which approach that of 
the fastest carbonic anhydrase, isozyme II. Heck et al. (1996) found that replacement of Tyr 
64, an inefficient proton shuttle, with a histidine in MCA V and removing a bulky residue at 
position 65 enhanced the maximal turnover for C0 2 hydration to values up to 10-fold over 
wild-type and 80-fold at physiological pH. Similar results have been obtained upon the 
replacements Lys 64 to His (Jewell et al., 1991) and Arg 67 to His (Ren et al., 1995) in the 
least efficient carbonic anhydrase, isozyme III, and with the replacement Asn 67 to His in 
isozyme II (Liang et al., 1993). 

75 



76 
The enhancements in catalysis caused by these insertions of histidine residues are 
observed in the steady state rate constant k cal for C0 2 hydration. These pH profiles of k cat 
can be described by a single ionization with a pK a near 7, suggesting the presence of the 
inserted histidine residue. There is a small or no effect on k^/K^ compared with wild-type 
in these mutants. Studies of pH dependencies and solvent hydrogen isotope effects have 
shown that these intramolecular proton transfers are rate-determining for maximal velocity 
and that these rate enhancements are specifically attributed to intramolecular proton transfer 
steps to His 64, just as in isozyme II (Silverman and Lindskog, 1988; Tu et al., 1989a). 

In isozyme II and III small buffers in solution, such as imidazole and derivatives, can 
act as proton donors and acceptors in the catalysis (Tu et al., 1989a; Tu et al., 1990). This 
is achieved in the mutant of isozyme II with the replacement of the predominant proton 
shuttle residue, His 64 to Ala (Tu et al., 1989a). These catalytically relevant buffer 
enhancements have been observed in steady state experiments and 18 exchange at chemical 
equilibrium and are saturable and consistent with proton transfer to the zinc-bound hydroxide 
(Tu et al., 1990, 1983; Paranawithana et al., 1990). The maximal rate constants for rate- 
limiting proton transfer from these buffers yield magnitudes similar to the amino acid 
counterpart, histidine. Therefore, the introduction of proton shuttles through site-directed 
mutagenesis or buffers in solution both are capable of achieving proton transfer rates nearly 
as rapid as in the most efficient carbonic anhydrase II. 

Various imidazole, pyridine, and morpholine buffers are used in this study as proton 
donors to enhance catalysis by a mutant of MCA V lacking a single predominant proton 
shuttle. The exchange of I8 between C0 2 and water was measured by mass spectrometry 






77 
to determine rate-limiting proton transfer from these buffers to the zinc-bound hydroxide in 
the dehydration direction of catalysis. The values of pK a of the buffers were ranged from 5.4 
to 8.6. Similar to previous intramolecular proton transfer studies in mutants of HCA III 
containing a histidine as a proton shuttle, the rate constants for proton transfer between the 
proton donor, buffer in this case, and zinc-bound hydroxide as acceptor are described in a free 
energy plot. Also similar is a curvature in this plot that is characteristic of fast and efficient 
proton transfers. Application of Marcus rate theory shows that this proton transfer has the 
small intrinsic energy barrier (near 0.8 kcal/mol) which is also characteristic of nonenzymic 
rapid proton transfer between nitrogen and oxygen acids and bases in solution. The Marcus 
parameters yield an observed overall energy barrier (near 10 kcal/mol), indicating that, similar 
to intramolecular counterparts in catalysis by carbonic anhydrase, there is a large involvement 
of energy requiring processes such as solvent reorganization or enzyme conformational 
change. This buffer enhancement study is interpreted in terms of the intramolecular 
counterparts in catalysis by isozymes of carbonic anhydrase. 

Materials and Methods 

Site-Specific Mutagenesis. Protein Expression and Purification 

The mutant MCA V Y64A/F65A was prepared by Dr. Minzhang Qian as described 
in Chapter 3 (page 53). Similar to all the MCA V mutants in my work (see Chapter 3), this 
mutant is a truncated form lacking the first 5 1 amino terminal residues and therefore, begins 
at Ser 22, in a sequence numbering scheme consistent with CA II. The expression and 
purification of this mutant is described in detail in Chapter 3 (page 55). 



78 
18 Exchange 

An Extrel EXM-200 mass spectrometer utilizing a membrane permeable to gases was 
used to measure the rate of exchange of 18 between species of CO, and water catalyzed by 
MCA V Y64A/F65A (equations 2-1 and 2-2, Chapter 2, page 27; Silverman, 1982). This 
18 0-exchange method is carried out at chemical equilibrium and can therefore be performed 
without buffers, a specific advantage in this study to determine proton transfer from buffers 
as proton shuttles. Solutions contained 25 mM total substrate ([C0 2 ] + [HC0 3 ']). Buffers 
were added only where indicated, and a total ionic strength of 0.2 M was maintained with 
Na 2 S0 4 at 25 °C. This method is described in detail in Chapter 2 on pages 27 through 29. 
Solvent hydrogen isotope effect experiments are as described on page 29 in Chapter 2. 

Results 

Method 1 : Saturation Effect of Buffers on R H 2c/[E] 

The effect of two buffers on R,, the rate of interconversion of C0 2 and HC0 3 ", and 
R H20 , the rate of release of 18 0-labeled water from the active site in catalysis by MCA V 
Y64A/F65A are shown in Figure 4-1 for 3,5-dimethyl pyridine and in Chapter 3 Figure 3-5 
for imidazole. Similar to other buffers of small size in carbonic anhydrase III, these two 
buffers enhance R^o^E] in a saturable manner for the mutant MCA V Y64A/F65A (Figure 
4-1 and 3-5; Tu et al., 1990, 1983; Paranawithana et al., 1990). This observed saturation is 
consistent with proton transfer from the buffers to the zinc-bound hydroxide in catalysis. The 
addition of each buffer listed in Table 4-1 to Y64A/F65A resulted in similar saturation plots 
for Rh2c/[E] a s shown for the example in Figure 4-1 . 



79 



GO 



H 



o 

is 



& 10 4 



c 

68 

p? 




10 20 30 40 50 60 70 
[protonated 3,5-DimethyI pyridine] mM 



Figure 4-1. The dependence of Rh2o/[ e ] (•) an d R i/[E] (□) as a function of the 
concentration of protonated 3,5-dimethyl pyridine at pH 6.3 and 25 °C. The total 
concentration of all species of C0 2 was 25 mM and the total ionic strength of solution was 
maintained at a minimum of 0.2 M by addition of NajSCV Rh2o/[ e 1 data are fit to equation 
4-1 yielding a value for k B app of (1 .7 ± 0. 1) x 10 4 s" 1 represented by the solid line. The dotted 
line is a fit to the same equation for Rh2o/[E] data points when corrected for the apparent 
inhibition manifested in R,/[E] yielding k B app of (2.5 ± 0. 1) x 10 4 s _l and a K^ 8 . K./ is listed 
in Table 4-1 . The data points of R,/[E] were applied to a least-squares fit, represented by the 
solid line, of catalytic velocity to the expression for competitive inhibition as a function of 
inhibitor concentration under the conditions that the total substrate concentration ([C0 2 ] + 
[HC0 3 '] ■ 25 mM) was much less than the apparent binding constant for total substrate, Kj ff s . 
The Kj was determined to be 176 ± 42 mM. 






to 

-a 






> 

< 
U 



< 

m 

o 
— 



vO 

>• 



c/i 

a: 

u 



m 



3 

CQ 

X> 

c 
u 

E 

CD 

o 

c 
CO 

-C 
c 



Q. 



<N 



O 

c 
o 
Q 

« 

a. 



•a 

c 

CO 






cO 
C 

W 

Cm 

O 
t/5 

c 

CO 

4—» 

c 
o 
U 



o 

B 

o 

a 



X> 
CO 

H 



3 



© © © © o o © © 

-H-H-H-H-H-H-H-H 

cs \ •— i cn i « cn ci cn — ' 



© O © d © 

-H -H -H -H -H 

in o in o »■« 

— O CN CN CN 



O 



-H 



■H 
VO </1 CN 

— o m 



Ov vo t~- 

■H -H -H 






-H 



o "* o m 

- N N M 



-H 
CN 



vo 

— vo in 

-H -H -H 

Ov m 00 



© © o" «-• d o ci © 

-H+l-H-H-H-H-H-H 
vomoot^iAiinTfO 
O— *-« m' ■* ' m m m 



CN CN CN CN 

©' © d d 

-H -H -H -H 

Ov cn rt- m 

^ _' _" (NJ 



WOvONTteOnh- CN CN •<r O 

7 •: o 6 6 6 ^ h odd — 



vo — t~~mr^mo\T 

ob od t— ' f* ' vc i vo i v~i in 



as ov r-- — c 
vo vo vo vd 







JU 




CD <U 








O 




C C 






u 


I 

-a 


V 


"9 ^ „, 






"5 

N 


o 

1 

-a 


>^ >>.E 






CO 

-a 


J 


O. Q.-0 




B 


J 




J 


P S 1 >^ 












CD CD — 




"5 
p. 




1 




E E £ 

9 9 >? 


u 

g 




2 


5 

i 


in 


1 


t in 2 




CN 


rt 


■— ' 


en' m m 


c- 



c 



■P "O T3 „ 

•c -c -c g 

a. a. D.-a 
">>*>, ">^ >, 

J= JS X! q] 
DUD- 

£ E E_£> 
S Q 5 1 
rt vo in «2 
CN (N (N (N 



CO ^ O "O <U g_ 00 JZ —< CN t*\ -rf 



-H 

<N 

00 



o r~~ 
«n d 

-H -H 
vo r- 



CN — 

d d 



rn CN 

r-' vo 



u 



^ 



N 



S5^ 



c« 



2 -H 
O — 



CO 



U 



^58 



D 



o 



E 

CO 



P _ ** 

U CO g 



<U 



CO 



SL-5 






<u 



CO 






a. 

CN 



CO 



CD 



c — 






CO 



pj 



o 



ffi 1 ^, — iT3 in ST3.CD 



CO 



3 Soil a 1 g 


















81 
The data of R^jo/fE] for 3,5-dimethyl pyridine and each buffer listed in Table 4-1 are 
described by equation 4-1, which assumes proton transfer from buffers, designated BH, to the 
zinc-bound hydroxide (Tu et al., 1990). 

IWtE] = k^tBHl/C*;/ + [BH]) + R^ (4-1) 

Here k B app is the maximal rate constant for the exchange of 18 to water that is enhanced by 
the buffers as proton donors and K eff B is an apparent binding constant of the buffer to the 
enzyme. [E] and [BH] are the concentrations of total enzyme and total buffer. R mo ° is the 
rate of release of 18 into solvent water at zero concentration of buffer and represents the 
contribution from the enzyme. The experimental values of R H20 /[E] determined by Method 
1 were fit to equation 4-1 using least-squares methods to determine k B app and K eff B . Listed in 
Table 4-1 are the values of K cff B for each buffer. 
Method 2: pH Dependence of R H2 C /[E] 

The pH dependence of R H2C /[E] catalyzed by MCA V Y64A/F65A was determined 
in the absence and presence of saturating concentrations of two buffers, 3,5-dimethyl pyridine 
and imidazole (Figure 4-2). A plot difference plot of Rh2c/[E] for the data in Figure 4-2 is 
shown in (Figure 4-3). The saturation concentration of each of these buffers was confirmed 
in the experiments described in Method 1 This analysis is similar to previous pH profiles of 
R H2c/[E] catalyzed by HCA III and mutants of HCA III in the presence of large 
concentrations of small buffers (Paranwithana et al., 1990; Tu et al., 1990; Jewell et al., 
1991). 

The rate constant for proton transfer from the donor group to the zinc-bound 
hydroxide and corresponding pK a values of the donor and acceptor were determined from 






82 







Figure 4-2. Variation with pH of Rh2c/[E], the proton-transfer dependent rate constant for 
the release from the enzyme of 18 0-labeled water, catalyzed by MCA V Y64A/F65 A in the 
(•) absence of buffer and in the presence of (□) 100 mM 3,5-dimethyl pyridine and (A) 100 
mM imidazole at 25 °C. For the data points of Y64A/F65A in the absence of buffer, the solid 
line represents a nonlinear least squares fit of equation 4-2 to the data with the pK a of the 
donor > 9 and acceptor = 6.9 ± 0.3 and the rate constant for proton transfer, k B = (3 .7 ± 0.4) 
x 10" 4 s' 1 . The total ionic strength of solution was maintained at 0.2 M by addition of NajSO,,; 
the total concentration of all species of C0 2 was 25 mM. 



83 




Figure 4-3. The difference in Rh2c/[E], the proton-transfer dependent rate constant for the 
release from the enzyme of 18 0-labeled water, between MCA V Y64A/F65 A in the absence 
and presence of (□) 100 mM 3,5-dimethyl pyridine and (A) 100 mM imidazole at 25 °C. 
Experimental conditions are described in the legend to Figure 4-2. The solid lines are 
nonlinear least-squares fits of equation 4-2 to the data with values of pK a for the proton 
donor, acceptor and k B given in Table 4-2. The fit to the data points for the difference 
between Y64A/F65A in the absence and presence of 100 mM 3,5-dimethyl pyridine at pH > 
7.0 was not considered. 



84 
the fit of the data in Figure 4-2 and 4-3 to equation 4-2 (this equation is described in Chapter 

2, equation 2-4, page 29) 

IW[E] = ^/{(l + VtHIXl + [HI/Ke)} (4-2) 

In equation 4-2, the value of R H2C /[E] can be interpreted in terms of the rate constant from 

a predominant donor group to the zinc-bound hydroxide (Silverman et al., 1993), in which 

k B is the rate constant for proton transfer to the zinc-bound hydroxide, K B is the ionization 

constant for the donor group and K E is the ionization constant of the zinc-bound water 

molecule. The values of k B and pK, for the donor and acceptors determined for the difference 

between Y64A/F65A and Y64A/F65A in the presence of saturating amounts of buffer are 

given in Table 4-2 and described in more detail in the following sections. 

In the absence of buffer, k B determined for Y64A/F65A contains a maximum near 4 

x 10 4 s" 1 at high pH (legend of Figure 4-2). This k B associated with high pH is also found for 

Y64A/F65A catalyzed in the presence of either 3,5-dimethyl pyridine or imidazole buffer 

(Figure 4-2). This can be interpreted as the contribution to k B from proton shuttles of high 

pK a on the enzyme; this topic is thoroughly discussed for catalysis by MCA V and mutants 

in Chapter 3. However, unique to the pH profiles with saturating amounts of buffer is the 

appearance of maxima near pH 6 to 7. For example, for the addition of 3,5-dimethyl pyridine 

to the pH profile of Y64A/F65A a maximum appears at pH 6.5 that was not present in its 

absence (Figure 4-2). By this argument, the maximum near pH 6.5 in the pH profile 

Y64A/F65A including 100 mM 3,5-dimethyl pyridine is due to the capacity of 3,5-dimethyl 

pyridine to act a proton donor. This suggests that the buffer 3,5-dimethyl pyridine is 

contributing to the rate of proton transfer to the zinc-bound hydroxide. This is supported by 



85 



Table 4-2: Maximal Values of k B with pK a Obtained from their pH Profiles for the 
Y64A/F65A Mutant of Murine Carbonic Anhydrase V, in the Absence and Presence of 
Buffer: Method 2. 

pK a pK a k B 

Buffer (Donor) (Acceptor) (x lO'V) 

Y64A/F65A + 

3, 5 -Dimethyl pyridine 6.2 ±0.1 6.9 ±0.1 2.5 ±0.8 

Y64A/F65A + 

Imidazole 6.4 ±0.1 7.5 7.2 ±1.6 

Note : All data were obtained by Method 2 with experimental conditions described in Figure 
4-3. 









86 
the difference between the pH profile for Rh2c/[ e ] f° r M CA V Y64A/F65 A in the presence 
and absence of 100 mM 3,5-dimethyl pyridine shown in Figure 4-3. This difference plot 
shows the bell-shaped pH dependence consistent with the addition of a single proton shuttle 
and corresponding to an additional proton transfer capacity (Rh2c/[E]) of (2.5 ± 0.8) x 10 5 
s" 1 with the presence of 3,5-dimethyl pyridine. Similar results were obtained upon addition 
of imidazole to Y64A/F65A (Figure 4-2 and 4-3; Table 4-2) 
pK a of the Donors and Acceptors Listed in Table 4-1 

pK, of the zinc-bound water . k^/K^ contains the rate constants for the steps in 
equation 1-1 (Chapter 1), the interconversion of CO, and HC0 3 " (Silverman and Lindskog, 
1988). The pH dependence of k ca /K m for the hydration of C0 2 is dependent on the ionization 
state of the zinc-bound water and therefore yields its pK a (Simonsson and Lindskog, 1982; 
Lindskog, 1983). The apparent pK a value of the zinc-bound water in the double mutant 
Y64A/F65A of MCA V was determined from the pH dependence of k^/K,,, obtained from 
18 0-exchange methods (Data not shown). A nonlinear least squares fit was applied to the 
data of k^JK^ and yielded a maximum at high pH and a single ionization with a pK, value of 
7. 1 ± 0.2. This value of pK a for the zinc-bound water agrees within experimental error with 
that determined by Method 2 for pH profiles of R H2C /[E] (Figure 4-2). R H 2c/[E] » s described 
by the ionization states of the proton donor and acceptor in the dehydration direction of 
catalysis and the pK a of the zinc-bound water determined by Method 2 is 6.9 ± 0.3. 
Therefore, the pK, of the zinc-bound water for each additional experiment upon which buffer 
is added to enhance catalysis was assigned the value of 7. 1 (Table 4-1). 



87 
As a control, this pK a value was compared to those determined by Method 2 in the 
experiment shown in Figure 4-3, where, saturating amounts of two buffers, 3,5-dimethyl 
pyridine and imidazole, were added to MCA V Y64A/F65A and Rh2c/[E] was determined. 
In each case, the pK a of the zinc-bound water determined from Rh2o/[E] was not greatly 
changed from that of the mutant MCA V Y64A/F65A determined in the absence of buffer 
(Table 4-2). This finding validates the assumptions given in Table 4-1 that the pK a of the 
zinc-bound water remains relatively unchanged upon addition of buffer to the enzyme. 

pK, of the Buffers. The pK, values of the buffers listed in Table 4-1 were determined 
by pH titration of each buffer in solution (see legend of Table 4-1). As described in the 
previous section the pH profile of MCA V Y64A/F65A was also determined in the presence 
of saturating amounts of two buffers, 3,5-dimethyl pyridine and imidazole by Method 2 
(Figure 4-2 and 4-3). For 3,5-dimethyl pyridine, the pK a determined in the active site of 
MCA V Y64A/F65A was found to be nearly identical to that determined from titration of the 
buffer in solution, within experimental error (Table 4- 1 and 4-2). Since there is a similarity 
of pK a values determined through pH titration of the buffers in solution to the pK a values 
determined in Method 2 in the pH profiles of R H20 /[E] for Y64A/F65A, for each buffer in 
Table 4-1 the pK a is determined by pH titration in solution 4 . However, an exception is the 
buffer imidazole that has its pK, for the donor decreased by 0.8 pH units in measurements of 



4 The addition of large amounts of buffer to the enzyme may change the charge distribution 
of the active site. In carbonic anhydrase this is possible because the active site is comprised 
of hydrophobic and hydrophilic sites. To avoid changes in the electrostatic potential of the 
active site, we held the ionic strength constant at 0.2 M in our buffer experiments. However, 
large concentrations of buffer may cause an increase in the ionic strength and thereby changes 
in pK a of the buffer or zinc-bound water. 



88 
R H2C /[E]; but as will be described in the coming sections imidazole behaves differently than 
the other buffers tested (Table 4-1 and 4-2). 
Determination of k B Values of Table 4-1 by Method 1 

The values of k B listed for the buffers in Table 4-1 were obtained by applying k B ipp , 
that was determined in Method 1, to equation 4-2. k B app represents the maximal rate constant 
for the exchange of 18 at saturating concentrations of total buffer. To determine the 
maximal rate constant from a protonated donor to an unprotonated acceptor for the 
dehydration direction of catalysis the value of k B app must be corrected for the concentration 
of protonated buffer and unprotonated enzyme as in equation 4-2. Equation 4-2 describes 
the pH dependence of R H2C /[E], the rate of release of 18 0-labeled water from the active site, 
which results in a bell-shaped curve for proton transfer from a proton donor and acceptor of 
equal pK,. The ionization constant for the zinc-bound water applied to equation 4-2 was 7. 1 
and the ionization constant of the buffers are listed in Table 4-1. 

The values of k B determined from Method 1, the buffer dependence of R^^E], in 
Figure 4-1 for 3,5-dimethyl pyridine, are close to the values of k B determined from Method 
2, pH profiles of MCA V Y64A/F65A in the presence of a saturating amounts of buffers 
(Table 4-2). The small difference in the k B value determined from the two methods for 3,5- 
dimethyl pyridine may be a function of buffer inhibition which is accounted for in Method 1, 
by direct measure of the inhibition in R, (Figure 4-2). Inhibition by these buffers is then 
corrected in our calculations of R H2C /[E] for Method 1 only (see legend to Figure 4-1). 
However this is not taken into account with the data obtained from Method 2, from pH 
profiles (Figure 4-2 and 4-3), therefore this may result in the values of k B being somewhat 












89 
smaller. The maximal value for R,/[E] in the pH profiles of Y64A/F65A in the presence of 
100 mM of either imidazole or 3,5-dimethyl pyridine is lower than in Y64A/F65A in the 
absence of buffer which may indicate inhibition as just described in the above sentences. 

The results obtained from Method 2 give similar values of k B to those determined 
through Method 1. Therefore, the values of k B for the other buffers of Table 4-1 were 
determined in a similar fashion, from Method 1, with these data for k B given in Table 4-1. 
There are two advantages for determination of k B by Method 1 . The first advantage is the 
direct measure and correction in k B for inhibition and the second advantage is that this method 
allows direct verification that the rate constant for proton transfer is determined at saturation. 
Solvent Hydrogen Isotope Effects 

The solvent hydrogen isotope effects (SHIE) observed for catalysis by MCA V 
Y64A/F65A in the presence of each of the buffers listed in Table 4-1 had a narrow range 
between 0.7 to 1.4 for k^/K,,, for CO, hydration. The SHIE determined for k B for 
Y64A/F65A in the presence of each buffer is also given in Table 4-1. Y64A/F65A in the 
presence of 3,5-dimethyl pyridine had the highest value of 3.4 for the SHIE consistent with 
rate-determining proton transfer. The lowest SHIE for k B was 1.5 which was determined for 
Y64A/F65A in the presence of 2,4 dimethyl pyridine (Table 4-1). The SHIE values listed for 
k B in Table 4- 1 are for total isotope effects from the enzyme and saturating buffer. 

Discussion 

Choice of Mutant and Buffers 

Previous work has described enhancement of proton transfer rates from buffers in 
solution by isozyme II (Tu et al., 1989a; 1990; Taoka et al., 1994) This study has extended 



90 
this previous work to buffers in solution in another CA, isozyme V and provides the most 
complete Marcus analysis of buffer catalysis by any enzyme (Table 4- 1 ). The choice of buffer 
was kept consistent with several criteria: a small size, a heterocyclic ring with nitrogen atoms, 
and only methyl substitutions on the ring. The mutant Y64A/F65A was chosen for these 
studies in order to compare its activation by buffers to the intramolecular proton transfer from 
His 64 in Y64H/F65A MCA V (Heck et al., 1996). These experiments have several other 
differences and advantages over the steady-state experiments of Taoka et al. 1994 that used 
five imidazole and pyridine type buffers to enhance catalysis in the H64A mutant of isoyzme 
II. First, my experiments offer a more extended pK a range using a series of buffers with 
values of the apparent pK a for the donor group ranging from approximately 5.4 for pyridine 
to 8.6 for morpholine. Second, this study contains a larger number of buffers to increase the 
accuracy of my results. And last, the 18 0-exchange technique was chosen to study the effects 
of buffer enhancements on catalysis. 18 exchange at chemical equilibrium offers the 
advantage that measurements of R H2C /[E] can be made in the absence of buffer. This allows 
a separation of the effects of intramolecular proton transfer rates from those of 
intermolecular (upon the addition of buffer), and it also allows the determination of the 
contribution of the addition of buffer and separately the contribution from the enzyme. 
Enhancement of Catalysis 

As depicted in equation 2-3 and 2-4, R H 2c/[E] measures a proton transfer dependent 
step of catalysis that is separate and distinct from the interconversion of C0 2 and HC0 3 ". The 
addition of each buffer listed in Table 4-1 to MCA V Y64A/F65A increased R mo /[E] in a 
saturable manner. This is consistent with the role of buffer as a second substrate and the 



91 
formation of buffer in an ES complex with the enzyme (Figure 4-1). Measurements of k B 
were determined at high concentrations of buffer, in the saturation region of a plot of R^fE] 
such as in Figure 4-1, to allow maximal binding of the substrate and to measure only the 
contribution of proton transfer from a bound site. This is in contrast to measurements at low 
concentrations of buffer where the proton transfer depends on a second-order process and 
our measurements are less precise. 

Several arguments suggest that the rate constant for proton transfer to the zinc-bound 
hydroxide, k B , are measuring the rate-determining step of catalysis. First, previous work from 
chemical equilibrium and steady state experiments, pH profiles, and SHIE described in 
Chapter 3 has demonstrated that the high values of k B and k cal found for this isozyme 
represent the rate-determining step in catalysis (Heck et al., 1994; 1996). Second, the proton 
transfer between zinc and buffer yields solvent hydrogen isotope effects for k B ranging from 
1 .5 to 3.4 (Table 4-1). This is consistent with rate-limiting proton transfer for k B . Values for 
solvent hydrogen isotope effects were determined to be near unity on k^/K,,, for the steps of 
the interconversion of C0 2 and HC0 3 \ a step separate and distinct from the proton transfer 
steps in catalysis. Therefore, the activation of R H20 /[E] upon increasing concentrations of 
buffer is a direct measure of the effect of buffer on proton transfer. The contribution from 
proton shuttle residues on the mutant is determined separately, from the dependence of 
R H20 /[E] on pH such as in Figure 4-1 and 4-2 at zero concentration of buffer. 
Bronsted Analysis 

The buffers studied in Table 4-1 were found to fall into three classes. The first class 
represents buffers "a" through "h" of Table 4-1 that are of the methyl-substituted imidazole, 



92 

pyridine and methyl-substituted pyridine at the 3, 4 and 5 positions, and morpholine type. 
The rate constants determined for proton transfer in MCA V Y64A/F65A in the presence of 
the buffers "a" through "h" listed in Table 4-1 correlated with the difference in pK, of the 
zinc-bound water as acceptor group and the buffer as donor in the dehydration direction of 
catalysis (Figure 4-4). Such Bronsted relationships are described by equation 4-3 (Bronsted 
andPederson, 1924; Kresge, 1973, 1975): 

log(k B ) = p[pK a (acceptor) - pK a (donor)] + constant (4-3) 

The slope P of a Bronsted plot provides analysis of the transition state structure in terms of 
the degree of charge transfer between reactants and products in the transition state. Proton 
transfer reactions involving carbon acids or bases usually yield slow rates and result in linear 
Bronsted relationships whereas fast proton transfer reactions between nitrogen and oxygen 
acids and bases are characterized by an observed variation in slope P over a relatively small 
range of pK a values (3 to 5 pK, units). This is curvature is also observed in Figure 4-4 for the 
buffers "a" through "h". 

The second class of buffers are also methyl-substituted pyridines, however these 
pyridines have a methyl substitution at the 2 or 6 position (designated "1" through "4" in 
Table 4-1). The rate constants for proton transfer, k B , from these buffers to the zinc-bound 
hydroxide showed greater scatter from the Bronsted plot than the previous class of buffers 
(Figure 4-4). An explanation for this observation is that methyl groups at carbons 2 or 6 may 
hinder the nitrogen general acid and base properties of substituted pyridines (Jencks, 1969), 
in contrast to the buffers "a" through "h". A linear fit to points "a" through "h" gives a low 
linear correlation coefficient (R 2 = 0.68). Finally, the buffer imidazole is in a class by itself 






93 



(ft 



3 




-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 

ApK s 



-a 



Figure 4-4. Dependence of the logarithm of k B (s"') on ApK, (the pK, of the zinc-bound water 
subtracted from the pK a of the donor group). Values of k B are listed in Table 4-1 and the 
experimental conditions are described in the legend of Table 4-1. The value of pK a for each 
buffer and the zinc-bound water are determined as described in Table 4-1 . The solid line is 
a nonlinear least-squares fit of equation 4-4 to the Marcus equation for points "a" through 
"h" This fit yielded values of AG* of 0.8 ± 0.5, w* of 9.8 ± 0.2 and w" of 8.2 ± 1.0. 



94 
in that it has the greatest rate constant for proton transfer than any buffer listed in Table 4-1 . 
Imidazole has been characterized in previous work as an exceptional acid and base catalyst 
(Jencks 1970; Scheiner and Yi, 1996). Its small size, similar pK a to the conjugate acid of the 
proton acceptor, zinc-bound hydroxide, and lack of steric hinderance or confinement in the 
active site may all be factors in the enhancements of proton transfer observed for Y64A/F65A 
in the presence of this buffer. 



Scheme 4-1 
^ ZnOH" 

' L_ 



H + -Buffer 



solvent, active-site 
reorganization 



. J^ZnOH" 



H + -Buffer 



AG* 



proton transfer 



„ ZnOH ? 



Buffer 



-\vP 



solvent, active-site 
reorganization 



. J ZnOH, 



Buffer 



Marcus Rate Theory 

The curvature in Bransted plots, similar to that found in Figure 4-4, can be interpreted 
through Marcus rate theory which provides an intrinsic energy barrier for proton transfer and 
two work terms that correspond to the energy required to align the reactants into the reaction 
complex for both forward and reverse directions (Marcus, 1968; Kresge, 1975). This is 
represented in Scheme 4-1 (Silverman et al., 1993) where w* is the energy required to arrange 



95 
the acceptor and donor groups and the water structure in the active site appropriate for the 
proton transfer in the dehydration direction (Silverman et al., 1993). The work term vf is the 
energy required for the same reorganization in the reverse direction of catalysis. AG : is then 
the intrinsic kinetic barrier with the appropriate active site orientation. The measured overall 
free energy for the reaction given by AG ■ W + AG R ° - W, where AG R ° is the standard free 
energy of reaction in the complex with appropriate orientation for proton transfer. 

The Marcus rate theory for proton transfer relates the observed overall activation 
barrier to proton transfer, AG*, equation 4-4, to the intrinsic energy barrier AG : : 

AG* = W + { 1 + AG R 74AG* } 2 AG* (4-4) 

The intrinsic energy barrier AG^ is the value of AG* when AG R °, the standard free energy of 
reaction with the required active site orientation (Schematic 4-1), is zero. AG J is then the 
barrier to reaction when the proton transfer is free of any thermodynamic constraints (Marcus, 
1968; Kresge, 1975). The slope of the Bronsted plot, P, equation 4-3 is identified in terms 
of the Marcus parameters as dAG J /AG° and provides the connection between the curvature 
of the Bronsted plot and the Marcus rate theory. 

A least squares fit of the Marcus equation, equation 4-4, was applied to the free 
energy plot constructed for buffers "a" through "h" of Table 4- 1 as proton donors and the 
zinc-bound hydroxide as proton acceptor (Figure 4-4). The values for the intrinsic energy 
barrier and work terms are given in Table 4-3 and legend to Figure 4-4. The intrinsic energy 
barrier for the proton transfer was found to be low with large work functions in both the 
forward and reverse directions (Table 4-3). Previous work involving intramolecular proton 
transfer rates from two sites in HCA III are indicated in Table 4-3. The values determined 



96 



Table 4-3: Marcus Theory Parameters for Proton Transfer in Isozymes of Carbonic 
Anhydrase. 

AG^ ~ w p 

System Proton Donor (kcal/mol) (kcal/mol) (kcal/mol) 

HCAIII His64 a 1.4 ±0.3 10.0 ±0.2 5. 9 ±1.1 

His67 b 1.3 ±0.3 10.9 ±0.1 5.9 ±1.1 

GluorAsp64 c 2.2 ±0.5 10.8 ±0.1 4.0 ±1.6 

MCAV Buffers' 1 0.8 ±0.5 10.0 ±0.2 8.2 ±1.0 

Nonenzymic Buffer to Buffer 6 2.0 3.0 

' Silverman et al., (1993) 

"Ren etal., (1995) 

c Tuetal, (1998) 

d Buffers are of the imidazole, pyridine, and morpholine type. The data were obtained by a 

least-squares fit of equation 4-4 to rate constants for proton transfer for points "a" to "h" 

given in Figure 4-4 and Table 4-1 . For this calculation AG : = -RT ln(hk B /kT) and AG° = RT 

e Kresge(1975) 






97 
from the contribution of buffer to catalysis are very similar in magnitude to the Marcus 
parameters describing intramolecular proton transfer from His 64 and His 67 to the zinc- 
bound hydroxide in mutants of HCA III. This low intrinsic energy barrier for proton transfer 
is very similar to nonenzymic proton transfer between nitrogen and oxygen acids and bases 
in solution. However, the large energy barrier present in the work terms of this and the 
previous studies is large compared to work terms of 2 kcal/mol for nonenzymatic reactions 
(Kresge, 1973; 1975). This has been interpreted as a requirement for water reorganization in 
the active site to provide a hydrogen-bonded water pathway for proton transfer between the 
proton shuttle and the zinc-bound hydroxide (Silverman et al., 1993). 
Conclusions 

It is significant that chemical rescue as quantitated by the Marcus parameters is so 
similar to proton transfer from His 64 or His 67. This suggests that the water structure 
appropriate for proton transfer is equally well formed to support catalysis by buffers or by His 
64, and is probably reflecting the great flexibility of many different water structures of 
approximately equal energy in the active site. The Marcus theory indicates that the required 
water structure and/or conformational change is infrequent, but when it occurs the proton 
shuttles in a facile manner similar to nonenzymic bimolecular proton transfers. 

My data serve another function in understanding proton transfer in carbonic 
anhydrase. The Marcus parameters obtained for His 64 and His 67 required the use of 
mutants of CA with replacements at position 198. These replacements of Phe 198 in HCA III 
altered the pK a of the zinc-bound water and allowed the construction of a Bronsted plot by 
varying the pK a of the proton acceptor in dehydration. All of the data of Figure 4-4 were 



98 
obtained with one enzyme. Therefore, the Marcus parameters that characterize proton 
transfer in this CA system are not predominately due to the changes in the structure at the 
active site of the mutants. 

The similarity of chemical rescue to proton transfer from His 64 is emphasized by the 
position of point "z", representing Y64H/F65A MCA V, on the same line of Figure 4-4 that 
represents chemical rescue. This supports the hypothesis that proton transfer involving His 
64 and buffers utilize a similar intervening water structure. Such a water structure is observed 
in the active site in the crystal structure of MCA V (Boriack-Sjodin et al., 1995), but of 
course this need not be the catalytically relevant water structure. 









CHAPTER 5 
DISCUSSION, CONCLUSIONS AND FUTURE DIRECTIONS 



The carbonic anhydrases catalyze the reversible hydration of C0 2 to bicarbonate; this 
reaction supports many physiological functions in mammals. The function of one carbonic 
anhydrase studied here, isozyme VII, is unknown, yet this isozyme has characteristics that 
suggest an important function. It is the most conserved of the mammalian carbonic 
anhydrases, is a very efficient isozyme, and has unique kinetic properties determined in this 
study. The second carbonic anhydrase investigated in this work, isozyme V, is suggested to 
have a metabolic function in providing bicarbonate to enzymes in the pathways of ureagenesis 
and gluconeogenesis (Dodgson, 1991). This isozyme has maximal turnover rates at high pH 
near 3 x 10 5 s" 1 . Isozyme V is located in mitochondria making its optimal catalytic activity 
at high pH ideal for its alkaline environment. However, the residue or residues that contribute 
to the maximal turnover at high pH was unknown before this study. 
Isozyme VII 

I have determined that isozyme VII has maximal turnover rates that place it among 
the fastest carbonic anhydrases, isozyme II and IV. This can be attributed, in part, to the role 
of histidine 64 as a proton shuttle in each of these isozymes. However, this work has defined 
some unique properties of isozyme VII that distinguish it from CA II and IV. 



99 



100 
Isozyme VII is the carbonic anhydrase most inhibited by the two sulfonamides, 
acetazolamide and ethoxzolamide. Previous work has determined that sulfonamides bind 
directly to the zinc and lie along the hydrophobic side of the active site cavity with 
interactions to the hydrogen bonded system of the zinc-bound hydroxide, Glu 106 and Thr 
199 (Figure 1-2; Eriksson et al., 1988b; reviewed in Liljas et al., 1994). The residues that 
define the hydrophobic pocket have been demonstrated to determine the degree of binding 
of the sulfonamides. Only two residues that are anticipated to define the hydrophobic 
environment of CA VII are not conserved in isozyme II. These are Leu 204 and Cys 206 in 
CA II which in murine CA VII are serines (Eriksson et al., 1988b; Hewett-Emmett and 
Tashian, 1996). Yet, a serine may be considered a conservative replacement of a cysteine 
residue. This may suggest that there are other differences in the active site of CA VII, such 
as subtle conformational effects that define its hydrophobic pocket compared to the other 
studied isozymes. A crystal structure of CA VII with an inhibitor bound in the active site 
would reveal the interactions of sulfonamides with the hydrogen bond network and with 
hydrophobic residues, and provide a comparison to sulfonamide binding to isozyme II 
(Eriksson et al., 1988b). 

His 64 was found to perturb the ionization state of the zinc-bound water upon pH 
measurements of the steady state constant k^/K^,. This is the first time in a wild- type CA 
that the influence of an active site residue on the ionization state of the zinc-bound water has 
been detected in k^/K^ and this defines another unique property of isozyme VII. This 
perturbation may indicate that there is a close interaction between residue 64 and the zinc 
which may facilitate the efficiency of His 64 as a proton shuttle in this isozyme. 



101 
Last, CA VII is unique in that its prominent proton shuttle residue His 64 is influenced 
by an interaction with the amino terminus. This is the first time a role for the amino terminus 
in catalysis by any carbonic anhydrase has been determined. The results of the CA VII 
truncation studies suggest several interpretations and functions of the amino-terminus. First, 
this interaction of His 64 with the amino terminus in isozyme VII may be a property shared 
with other isozymes of CA that also contain His 64 as a functional proton shuttle. Aronsson 
et al. (1995) constructed a series of truncation mutants in isozyme II and obtained a decreased 
value of overall C0 2 hydration activity with a 20 residue amino-terminal truncation mutant 
when compared to wild-type. Second, His 64 may be in a close environment with the amino 
terminus in isozyme VII. Crystal structures of isozyme II have shown that its proton shuttle, 
His 64, can assume two conformations "in" and "out" which may be catalytically relevant for 
the proton shuttle function in transferring protons between the zinc and buffer in solution 
(Nair and Christianson, 1991). The amino terminus in isozyme VII may force such 
conformations (by steric or electrostatic effects) that are required for proton transfer from His 
64. Finally, a proton relay mechanism from the active site to solution may extend beyond His 
64 and involve many histidine residues in the amino terminus. This is supported by the finding 
in Chapter 2 that truncation of a H64A mutant of CA VII further removes proton transfer 
capacity beyond that of the full-length H64A mutant. In murine CA VII the amino terminus 
has 3 histidine residues at positions 3, 4, and 17 based on the numbering scheme of CA I 
(Figure 2-1). 






102 
Isozyme V 

For isozyme VII, the pH dependence of k cal is described by two ionizations. One 
ionization is near pH 6 and is associated with the histidine at position 64. The other 
ionization at pH greater than 8 is of an unknown origin (Earnhardt et al., 1998a). Heck et al. 
(1994) demonstrated that murine isozyme V has maximal turnover rates at high with the pH 
profile of k^ dependent on one or more ionizations of pK a near 9. However, the residue(s) 
that contribute to the maximal turnover at high pH in this isozyme had not been identified at 
that time. The pH profiles of k ca , for at least three other isozymes in the a class, CA II, III 
and IV, also demonstrate a dependence on ionizations at high pH which cannot be attributed 
to His 64 (Silverman et al., 1998). This study determined the residues that contribute to the 
ionization in k cal and that support the catalytic activity of MCA V. These residues include 
a number of lysines and tyrosines located around the rim of the active-site cavity. Site- 
directed mutagenesis of Lys 91 and Tyr 131 to alanine residues resulted in a 50% decrease 
in k cat for hydration. A quadruple mutant with two additional suggested proton transfer 
residues (Tyr 64 and Lys 132) replaced to alanine results in a decrease in the catalytic 
turnover to 10% of wild-type. These findings suggest that MCA V can sustain proton 
transfer from the zinc to residues distributed throughout its active site and that position 64 
is not the only site from which proton transfer can occur in the active site environment of CA 
V. It is possible to consider this observation of shuttle groups of pK a near 9 in CA V and 
other isozymes as due to an accumulation of basic amino acids occurring near the active site 
cavity. Each of these isozymes (CA II, III, IV, and VII) contains a lysine at positions 
169/170 and 127 and, a tyrosine residue at 194. In addition, many tyrosine or lysine residues 



103 
are located at the mouth of the active site cavity or are present at positions 91 and 132 in 
most of these isozymes (Boriack-Sjodin, 1995; Hewett-Emmett and Tashian, 1996). 

Thus, in MCA V and likely in other isozymes of the a class of the carbonic anhydrases 
there are multiple proton transfers from basic residues contributing to the overall catalytic 
efficiency of catalysis. This rinding is in sharp contrast to the accepted mechanism of proton 
transfer for the well-studied isozyme II, the activity of which was previously described as due 
to a single predominant proton shuttle (Silverman and Lindskog, 1988; Tu et al., 1989a). In 
fact, this is the first evidence involving sites for proton transfer other than His 64, and the first 
evidence for lysine and tyrosine residues as proton shuttles in a carbonic anhydrase. And 
finally, this is the first quantitative evidence for multiple proton transfer in a carbonic 
anhydrase. 
Buffers in Catalysis 

Chemical rescue refers to the activation of a mutant enzyme lacking a critical amino 
acid residue by replacing that residue with an analogous reagent from solution. MCA V lacks 
a single predominant proton shuttle. In a chemical rescue experiment, buffers of the imidazole 
pyridine and morpholine type enhanced rate-determining proton transfer steps in a mutant of 
MCA V, Y64A/F65A to values similar to that of the intramolecular counterpart, His 64, in 
the mutant Y64H/F65A. The rate constants for proton transfer from these buffers to the zinc- 
bound hydroxide in catalysis of MCA V correlated with the difference in acid and base 
strength of the catalysts. Application of Marcus rate theory to these data determined that 
buffers behave similarly to their intramolecular counterparts for proton shuttling in enzymatic 
reactions, yielding low energy barriers to proton transfer (~ 2 kcal/mol) and large energy 



104 
barriers (~ 10 kcal/mol) for the work required prior to proton transfer to orient donor and 
acceptor and form the intervening water bridge that is necessary for proton transfer. 

This study has provided the most complete investigation of chemical rescue of any 
carbonic anhydrase and has determined that chemical rescue is similar to proton transfer from 
His 64 or His 67 as quantitated by the Marcus parameters. The energy barriers, described by 
the Marcus theory, indicate that the required water structure and/or conformational change 
for proton transfer is not frequently present, but when it occurs the proton shuttles in a 
manner similar to nonenzymic bimolecular proton transfers. This suggests that the water 
structure appropriate for proton transfer is equally well formed to support catalysis by buffers 
or by His 64. This may reflect the flexibility of many different water structures of 
approximately equal energy in the active site. Also, determined from this buffer work, the 
Marcus parameters of intramolecular proton transfer obtained from previous studies are 
apparently not influenced by changes at residue 198. As described in Chapter 4, the Marcus 
parameters for His 64 and His 67 required the use of mutants of CA with replacements at 
position 198 to alter the pK a of the zinc-bound water and allowed the construction of a 
Bronsted plot. The similarity of the Marcus parameters between the isozyme III work using 
mutants and this study with buffers emphasizes that replacements at position 198 are not 
influencing the Marcus parameters. 
Conclusions 

The topic of this dissertation has focused on the contribution of three modes of proton 
transfer in the carbonic anhydrases: near intramolecular, distant intramolecular, and 
intermolecular. A quantitative and comparative analysis of the rate constants for proton 



105 
transfer from histidine, tyrosine, and lysine residues and external buffers as proton shuttles in 
catalysis has been provided. Of each of these proton shuttles, the inherently most efficient is 
that of free imidazole buffer in solution and histidine residues in the active site. These results 
can be explained by the optimal distance of these proton shuttles from the zinc-bound water 
with imidazole buffers lacking distance constraints and His 64 located at a close spatial 
position in the active site. Other factors influencing the efficiency of these various proton 
shuttles include the distance to nearby interfering groups and the difference in pK a between 
the proton donor and acceptor groups. These studies focused on two isozymes of carbonic 
anhydrase and the interpretation has been expanded to include other isozymes. It is hoped 
that this analysis of proton transfer in the carbonic anhydrases may be applied to more 
complex proton transfer systems. 
Future Work 

The role of His 64 as a proton shuttle in catalysis by the most efficient isozymes of 
carbonic anhydrase, now including CA VII, has been established. However, the present work 
on CA VII has indicated that very little is known about the influence on catalysis of the amino 
terminus. Now that the wild-type and the H64A mutant of C A VII exists in an expression 
vector that is easily manipulated, it will be possible to determine if the three histidine residues 
located on the amino terminus contribute in a proton relay mechanism with His 64. Site- 
directed mutagenesis to produce single alanine mutants of these three histidine residues and 
double alanine mutants of each of the three histidines with His 64 will allow kinetic analysis 
of the properties of these mutants. Similar analysis as that described in Chapter 3 for the 
alanine mutants of CA V will allow the identification of any cooperative effects of histidine 



106 
residues on catalysis. A crystal structure of the native and truncated forms of CA VII will 
be useful for comparisons of conformations of His 64 in the presence and absence of the 23 
residue amino terminus and will be useful for identification of potential proton shuttles located 
on the amino-terminus. 

The intrinsic kinetic barrier, which defines the energy required for proton transfer, for 
intra or intermolecular proton shuttles such as histidine or imidazole type buffers in catalysis 
is similar in magnitude to that of proton transfer between nitrogen and oxygen acids and bases 
for bimolecular proton transfer in solution (2 kcal/mol; Silverman et al., 1993). The similarity 
of the intrinsic barrier determined from these enzymic and nonenzymic systems is understood 
because this barrier is interpreting pure kinetic effects. However, the energy required to 
orient the system and/or water for proton transfer is large in carbonic anhydrase, 10 kcal/mol, 
if it is compared to that of proton transfer between two buffers in solution (3 kcal/mol). The 
significance of the large work term in enzymatic reactions has not been clearly established. 
We have a very defined system in the isozymes of carbonic anhydrase that allows the pK a of 
the proton donor or acceptor to be varied to obtain Bronsted relations and determine the 
Marcus parameters for catalysis. With this system we can investigate what factors contribute 
to this large work term and determine how the work term is affected upon changes in the 
geometry of the active site, the degree of hydrophobicity of the active site and consequently 
changes in the solvent organization, and also from distance requirements for proton transfer. 
For example, buffers are an unimpeded source of proton shuttles and distance requirements 
most likely were not a factor in this present study of buffer enhancements on catalysis. 
Similarly, proton shuttle residues placed at position 64 and 67 are at a close distance to the 



107 
zinc. In addition, this work showed catalysis is sustained by proton transfer from more distant 
sites than position 64, with decreases in catalysis of 50% upon removal of the most effective 
proton shuttles in isozyme V. Therefore, positioning proton shuttles at greater distances from 
the zinc by site-directed mutagenesis with corresponding mutations at position 198 in isozyme 
III would allow the Marcus parameters to be determined. Any change observed in the 
Marcus parameters will aid in the interpretation of the large work term and aid in the 
identification of factors that are required in an enzyme active site for proton transfer. 

Last, the low intrinsic kinetic barrier for proton transfer determined for the carbonic 
anhydrases may be indicative of hydrogen tunneling. Experiments that are necessary to define 
hydrogen tunneling in an enzyme include temperature and solvent hydrogen isotope effects 
studies. The chemical rescue experiments defined in this study would allow the least 
complicated method to perform such hydrogen tunneling experiments. Temperatures at 
extreme degrees require stable enzymes and therefore, varying buffers as proton shuttles in 
one enzyme limits instability problems that are introduced in mutants. All of these proposed 
experiments should expand our current understanding of proton transfer and more specifically 
the role of proton shuttles and the role of the active site water in catalysis. 



REFERENCES 

Agmon, N. (1995) Chem. Phys. Lett. 244, 456-462. 

Alexander, R. S., Nair, S. K., and Christianson, D. W. (1991) Biochemistry 30, 1 1064-1 1072. 

Aronsson, G., Martensson, L. -G , Carlsson, U., and Jonsson, B. -H. (1995) Biochemistry 34, 
2153-2162 

Behravan, G., Jonasson, P., Jonsson, B. -H. and Lindskog, S. (1991) Eur. J. Biochem. 198, 
589-592. 

Boriack-Sjodin, P. A., Heck, R. W., Laipis, P. J., Silverman, D. N, and Christianson, D. W. 
(1995) Proc. Natl. Acad. Sci. USA 92, 10949-10953. 

Bronsted, J. N. and Pedersen, K. (1924) Z. Phys. Chem. 108, 185. 

Chen, X., Tu, C. K., LoGrasso, P. V., Laipis, P. J., and Silverman, D. N. (1993) 
Biochemistry, 32, 7861-7865. 

Christianson, D. W., and Fierke, C. A. (1996) Ace. Chem. Res. 29, 331-339. 

Dodgson, S. J. (1991) in The Carbonic Anhydrases (Dodgson, S. J., Tashian, R. E , Gros, 
G. and Carter, N. D , Eds.) pp 297-306, Plenum Press, New York. 

Earnhardt, J. N., Qian, M. Z., Tu, C. K , Lakkis, M. M, Bergenhem, N. C. H., Tashian, R. 
E., Laipis, P. J., and Silverman, D. N. (1998a) Biochemistry 37, 10837-10845. 

Earnhardt, J. N, Qian, M. Z , Tu, C. K., Laipis, P. J., and Silverman, D. N. (1998b) 
Biochemistry 36, 15758-15764. 

Earnhardt, J. N., Wright, S. K., Qian, M. Z., Tu, C. K., Laipis, P. J., Viola, R. E., and 
Silverman, D. N. (1998c) Arch. Biochem. Biophys. submitted for publication. 

Eriksson, A. E., Jones, T. A, and Liljas, A. (1986) In Zinc Enzymes (Bertini, I., Luchinat, C, 
Maret, W., and Zeppezauer, M., Eds) pp 317-328, Birkhauser, Boston. 

Eriksson, A. E., Kylsten, P. M., Jones, T. A., and Liljas, A. (1988a) Proteins: Struct. Func. 
and Gen. 4,274-282. 

108 



109 



Eriksson, A. E, Kylsten, P. M, Jones, T. A., and Liljas, A. (1988b) Proteins: Struct. Func. 
and Gen. 4, 283-293. 

Eriksson, A. E. and Liljas, A. (1993) Proteins Stntc. Fund. Genet. 16, 29-42. 

Feldstein, J. B., and Silverman, D. N. (1984) J. Biol. Chem. 259, 5447-5453. 

Fierke, C. A., Calderone, T. L., and Krebs, J. F. (1991) Biochemistry 30, 1 1054-1 1063. 

Forsman, C, Behravan, G, Jonsson, B. -H., Liang, Z. W., Lindskog, S., Ren, X., Sandstrom, 
J., and Wallgren, K. (1988) FEBS Lett . 229, 360-362. 

Guan, K. L., and Dixon, J. E. (1991) Anal. Biochem. 192, 262-267. 

Hakansson, K„ Carlsson, M., Svensson, L. A, and Liljas, A. (1992) J. Mol. Biol. 227, 1 192- 
1204. 

Heck, R.W., Boriack-Sjodin, P. A., Qian, M. Z., Tu, C. K., Christianson, D. W , Laipis, P. 
J., and Silverman, D. N. (1996) Biochemistry 35, 1 1605-1 1611. 

Heck, R. W., Tanhauser, S. M., Manda, R., Tu, C. K., Laipis, P. J., and Silverman, D. N. 
(1994) J. Biol. Chem. 269, 24742-24746. 

Hewett-Emmett, D., and Tashian, R. E. (1996) Mol. Phylogenet. Evol. 5, 50-77. 

Hurt, J. D., Tu, C. K., Laipis, P. J., and Silverman, D. N. (1997) J. Biol. Chem. 272, 13512- 
13518. 

Jackman, J. E., Merz, K. M. and Fierke, C. A. (1996) Biochemistry 35, 16421-16428. 

Jencks, W. P. (1970) Biochem. Soc. Symp. 31, 59-80. 

Jencks, W. P. (1969) in Catalysis in Chemistry and Enzymology, pp 96-97. McGraw-Hill 
Book Co., New York. 

Jewell, D. A., Tu, C. K., Paranawithana, S. R., Tanhauser, S. M., LoGrasso, P. V., Laipis, 
P. J. and Silverman, D. N. (1991) Biochemistry 30, 1484-1490. 

Jonsson, B. -H., Steiner, H. Lindskog, S. (1976) FEBS Lett. 64, 310-314. 

Kannan, K. K., Notstrand, B., Fridborg, K., Lovgren, S., Ohlsson, A., and Petef, M. (1975) 
Proc. Natl. Acad. Sci. USA 72, 51-55. 





















110 

Khalifah, R. G. (1973) Proc. Natl. Acad. Sci. USA 70, 1986-1989. 
Khalifah, R. G. (1971) J. Biol. Chem. 246, 2561-2573. 

Khalifah, R. G, Stader, D. J., Bryant, S. H., and Gibson, S. M. (1977) Biochemistry 16, 

2241-2247. 

Krebs, J. F., Ippolito, J. A., Christianson, D. W , and Fierke, C. A. (1993a) J. Biol. Chem. 
268, 27458-27466. 

Krebs, J. F., Rana, F., Dluhy, R A, and Fierke, C. A. (1993b) Biochemistry, 32, 4496-4505. 

Kresge, A. J. (1975) Ace. Chem. Res. 8, 354-360. 

Kresge, A. J. (1973) Chem. Soc. Rev. 2, 475-503. 

Kunkel, T. A. (1985) Proc. Natl. Acad. Sci. USA. 82, 488-492. 

Kunkel, T. A., Roberts, J. D., and Zakour, R. A. (1987) Methods Enzymol. 154, 367- 
382. 

Lakkis, M. M., Bergenhem, N. C. H., and Tashian, R. E. (1996) Biochem. Biophys. Res. 
Commun. 226, 268-272. 

Lakkis, M. M., O'Shea, K. S. and Tashian, R. E. (1997)7. Histochem. Cytochem. 45, 657- 
662. 

Liang, Z., Jonsson, B. -H., and Lindskog, S. (1993b) Biochim. Biophys. Acta. 1203, 142-146. 

Liang, J. Y. and Lipscomb, W. N. (1988) Biochemistry, 27, 8676-8682. 

Liang, Z., Xue, Y., Behravan, G, Jonsson, B. -H., and Lindskog, S. (1993a) Eur. J. 
Biochem. 211,821-827. 

Liljas, A, Hikansson, K, Jonsson, B. -H., and Xue, Y. (1994) Eur. J. Biochem. 219, 1-10. 

Lindskog, S. (1997) Pharm. Ther. 74, 1-20. 

Lindskog, S. (1984) J. Mol. Catalysis 23, 357-368. 

Lindskog, S. and Coleman, J. E. (1973) Proc. Natl. Acad. Sci. USA 70, 2505-2508. 



Ill 

Ling, B., Bergenhem, N. C. H., Dodgson, S. J., Forster, R. E., and Tashian, R. E. (1994) 
Isozyme Bull. 27, 47. 

Ling, B., Bergenhem, N. C.H., and Tashian, R. E. (1995) Isozyme Bull. 28, 32. 

LoGrasso, P. V., Tu, C. K., Chen, X., Taoka, S., Laipis, P. J. and Silverman, D. N. (1991) 
Biochemistry 30, 8463-8470. 

LoGrasso, P. V., Tu, C. K., Jewell, D. A., Wynns, G. C, Laipis, P. J. and Silverman, D. N. 
(1993) Biochemistry 32, 5786-5791. 

Marcus, R. A. (1968) J. Phys. Chem. 72, 891-899. 

Maren, T. H. (1997) J. Exp. Zool. 279, 490-497. 

Merz, K. M. (1990) J. Mol. Biol. 214, 799-802. 

Mildvan, A. S., Weber, D. S„ and Kuliopulos, A. (1992) Arch. Biochem. Biophys. 294, 327- 
340 

Montgomery, J. C, Venta, P. J., Eddy, R. L., Fukushima, Y. S., Shows, T. B., and Tashian, 
R. E. (1991) Genomics 1 1, 835-848. 

Nagao, Y., Platero, J. S., Waheed, A., and Sly, W. S. (1993) Proc. Natl. Acad. Sci. USA 90, 
7623-7627. 

Nair, S. K. and Christianson, D. W. (1991)7. Am. Chem. Soc. 1 13, 9455-9458. 

Paranawithana, S. R, Tu, C. K., Laipis, P. J., and Silverman, D. N. (1990) J. Biol. Chem. 
265, 22270-22274. 

Pocker, Y. and Bjorkquist, D. W. (1976) J. Am. Chem. Soc. 99, 6537-6543. 

Prabhananda, B. S., Rittger, E., and Grell, E. (1987) Biophys. Chem. 26, 217-224. 

Ren, X., Tu, C. K., Laipis, P. J., and Silverman, D. N. (1995) Biochemistry 34, 8492-8498. 

Rowlett, R. S. (1984)./. Prot. Chem 3, 369-393. 

Rowlett, R S. and Silverman, D. N. (1982) J. Amer. Chem. Soc. 104, 6737-6741. 

Scheiner, S. and Yi, M. (1996) J. Phys. Chem. 100, 9235-9241. 



112 

Scolnick, L. R, and Christianson, D. W. (1996) Biochemistry 35, 16429-16434. 

Silverman, D. N. (1982) Methods inEnzymology 87, 732-752. 

Silverman, D. N., Earnhardt, J. N., Wadhwa, N. R , Tu, C. K., Qian, M. Z. and Laipis, P. J. 
(1998) manuscript in preparation. 

Silverman, D. N., and Lindskog, S. (1988) Ace. Chem. Res. 21, 30-36. 

Silverman, D. N., Tu, C. K., Chen, X., Tanhauser, S. M., Kresge, A. J., and Laipis, P. J. 
(1993) Biochemistry 32, 10757-10762. 

Silverman, D. N. and Tu, C. K. (1975) J. Amer. Chem. Soc. 97, 2263-2269. 

Simonsson, I., Jonsson, B. -H., and Lindskog, S. (1982) Eur. J. Biochem. 123, 29-36. 

Simonsson, I., Jonsson, B. -H., and Lindskog, S. (1979) Eur. J. Biochem. 93, 409-417. 

Sly, W. S. and Hu, P. Y. (1995) Aram. Rev. Biochem. 64, 375-401. 

Smith, D. B., and Johnson, K. J. (1988) Gene 67, 31-40. 

Stams, T., Chen, Y., Boriack-Sjodin, P. A., Hurt, J. D., Liao, J., May, J. A., Dean, T., Laipis, 
P., Silverman, D. N., Christianson, D. W. (1998) Protein Sci. 7, 556-563. 

Stams, T., Nair, S. K., Okuyama, T., Waheed, A, Sly, W. S., and Christianson, D. W. (1996) 
Proc. Natl. Acad Sci. USA 93, 13589-13594. 

Steiner, H., Jonsson, B. -H., and Lindskog, S. (\915)Eur. J. Biochem. 59, 253-259. 

Studier, F. W., Rosenberg, A. H., Dunn, J. J., and Dubendorff, J. W. (1990) Methods 
Enzymol. 185, 60-89. 

Tanhauser, S. M., Jewell, D. A, Tu, C. K., Silverman, D. N., and Laipis, P. J. (1992) Gene 
117, 113-117. 

Taoka, S., Tu, C. K., Kistler, K. A, and Silverman, D. N. (1994) J. Biol. Chem. 269, 17988- 
17992. 

Tipton, K. F., and Dixon, H. B. F. (1979) Methods in Enzymol. 63, 183-234. 

Tu, C. K., Paranawithana, S. R., Jewell, D. A., Tanhauser, S. M., LoGrasso, P. V., Wynns, 
G. C, Laipis, P. J., and Silverman, D. N. (1990) Biochemistry, 29, 6400-6405. 



113 

Tu, C. K., Qian, M. Z., Earnhardt, J. N., Laipis, P. J., and Silverman, D. N. (1998) 
Biophysical J. 74, 3 1 82-3 1 89. 

Tu, C. K., Sanyal, G, Wynns, G. C, and Silverman, D. N. (1983) J. Biol. Chem. 258, 8867- 
8871. 

Tu, C. K., Silverman, D. N., Forsman, C, Jonsson, B. -H., and Lindskog, S. (1989a) 
Biochemistry 28, 7913-7918. 

Tu, C. K., Thomas, H. G, Wynns, G C, and Silverman, D. N. (1986) J. Biol. Chem. 261, 
10100-10103. 

Tu, C. K., Wynns, G C, and Silverman, D. N. (1989b) J. Biol. Chem. 264, 12389-12393. 

Tu, C. K., Wynns. G. C, and Silverman, D. N. (1981)./. Biol. Chem. 256, 9466-9470. 

Venkatasubban, K. S. and Silverman, D. N. (1980) Biochemistry, 19, 4984-4989. 

Verpoorte, J. A., Mehta, S., and Edsale, J. T. (1967) J. Biol. Chem. 242, 4221-4229. 

Wells, J. W., Kandel, S. I., Kandel, M, and Gornall, A. G. (1975)7. Biol. Chem. 250, 3522- 
3530. 

Xue, Y., Liljas, A., Jonsson, B. -H., and Lindskog, S. (1993) Proteins Strttc. Fund. Genet. 
17, 93-106. 






BIOGRAPHICAL SKETCH 

Nicole Earnhardt was bom on September 1 1, 1970 in Raleigh, North Carolina where 
she remained through high school and college. In the fall of 1988, after graduating high 
school, she attended North Carolina State University where she worked part time in a maize 
molecular biology laboratory on undergraduate research projects. She received her Bachelor 
of Science in Biological Sciences and a minor in Business Management in December of 1992. 
In one more semester of study, she completed a Bachelor of Science in Biochemistry in the 
spring of 1993. She then worked the following summer as a laboratory teaching assistant for 
a series of biotechnology courses at North Carolina State University. In the fall of 1993, she 
moved to Gainesville, Florida to begin graduate school in the Department of Biochemistry 
and Molecular Biology at the University of Florida. At the University of Florida she began 
her doctoral research with David Silverman, a professor in the Department of Pharmacology 
in the College of Medicine, where she studies proton transfer in the carbonic anhydrases. She 
intends to receive her Ph.D. from the University of Florida in the fall of 1998. After that she 
will continue her research career at the University of Florida where she will be a postdoctoral 
fellow in the laboratories of Drs. Harry Nick and Doug Anderson in the Department of 
Neuroscience. Her postdoctoral studies will include transcriptional regulation of putative 
inflammatory response genes in post spinal cord trauma. Her long term goals are to use the 
vast array of techniques and experiences she has gathered as a graduate student and 
postdoctoral fellow to manage a scientific laboratory in industrial science. 

114 



I certify that I have read this study and that in my opinion it conforms to acceptable 
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation 
for the degree of Doctor of Philosophy. 

/Q9u*J^/3Ub»* — 
David N. Silverman, Chair 
Professor of Pharmacology and 

Therapeutics 

I certify that I have read this study and that in my opinion it conforms to acceptable 
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation 
for the degree of Doctor of Philosophy. A A A 

. IL 

Philip J. Laidis/ 
Professor of Biochemistry and 
Molecular Biology 




I certify that I have read this study and that in my opinion it conforms to acceptable 
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation 
for the degree of Doctor of Philosophy. 



C&esW^J). 



Ben M. Dunn 
Distinguished Professor of 
Biochemistry and Molecular Biology 



I certify that I have read this study and that in my opinion it conforms to acceptable 
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation 
for the degree of Doctor of Philosophy. 

Thomas H. Mareci 

Associate Professor of Biochemistry and 
Molecular Biology 

I certify that I have read this study and that in my opinion it conforms to acceptable 
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation 
for the degree of Doctor of Philosophy. 




Harry S. 
Professor* -df Biochemistry and 
Molecular Biology 



I certify that I have read this study and that in my opinion it conforms to acceptable 
standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation 
for the degree of Doctor of Philosophy. 




Joh D. Stewart 

Assistant Professor of Chemistry 



This dissertation was submitted to the Graduate Faculty of the College of Medicine and 
to the Graduate School and was accepted as partial fulfillment of the requirements for the 
degree of Doctor of Philosophy. 



December, 1998 



/Ci^b^L Cla^j^-4- 

Pi iii CMea (irMinlii.uw'ifcj 



















UNIVERSITY OF FLORIDA 



3 1262 08555 3138